DOCTORAL THESIS

UNIVERSIDAD DE LA SABANA

SCHOOL OF ENGINEERING

DOCTORADO EN LOGISTICA Y GESTION DE CADENAS DE SUMINISTROS

Publicly Defended On April 2026, by:

Fernando Enrique Mantilla Patiño Industry 4.0 and Production Optimization in Ecuador’s Rose Export Sector: An integrated Approach

Tutors:

Mejía- Delgadillo, Gonzalo Associate Professor, Universidad de la Sabana, Colombia Supervisor Sarmiento-Vásquez, Alfonso Associate Professor, Universidad de la Sabana, Colombia Co-Supervisor Moreno-Camacho, Carlos

Co-Supervisor

Acknowledgements

In this section, I wish to express my sincere and profound gratitude to all those who supported me and contributed to the completion of this thesis. This work would not have been possible without their encouragement, guidance, and generosity. First and foremost, I am deeply grateful to my family for their unwavering support throughout this journey. I owe special thanks to my wife, Clara Lucía, for her patience, understanding, and sacrifice, particularly for the many hours I was unable to dedicate to our personal life due to my academic commitments.

I would also like to extend my heartfelt appreciation to my thesis supervisors, Gonzalo Mejía, Carlos Moreno, and Alfonso Sarmiento. Their continuous support, insightful guidance, and invaluable academic contributions were fundamental to the development and completion of this research.

Likewise, I wish to convey my sincere appreciation to the examiners of this thesis, for their time, methodological rigor, and scholarly insights. Their perceptive observations and constructive critiques significantly enhanced the refinement and academic quality of this work.

I am equally thankful to Universidad de La Sabana for providing me with the opportunity to pursue my doctoral studies. The academic environment, knowledge, and experiences I gained have been profoundly enriching. I also extend my gratitude to my fellow doctoral students and professors, whose guidance and shared expertise greatly contributed to my intellectual growth.

Finally, I express my deep gratitude to Falcon Farms de Ecuador S.A., where I currently work, for its institutional support. I am particularly indebted to my supervisor, Jairo Rengifo, whose leadership, trust, and exceptional human qualities have been a constant source of inspiration. His support played a decisive role in enabling the successful completion of this academic endeavor.

I am sincerely grateful to all who, in one way or another, contributed to this achievement. Their support has been invaluable both to this research and to my personal and professional development.

| iv

Table of contents

Acknowledgements .................................................................................................................................. iii Table of contents ...................................................................................................................................... iv List of Tables ............................................................................................................................................ vi List of Figures ......................................................................................................................................... vii Abstract .................................................................................................................................................. viii

1.

Preliminaries .................................................................................................................................... 9

1.1

Introduction ...................................................................................................................................... 9

1.2

Research Objective ...................................................................................................................... 100

1.3

Structure and Organization of the thesis ...................................................................................... 100

1.4

Overview of publication outputs .................................................................................................. 111

2.

The Role of Industry 4.0 in the Export Flower Industry ................................................................ 13

2.1

Introduction .................................................................................................................................. 144

2.2.1

Systematic Literature Review ...................................................................................................... 211

2.2.2

Exploratory surveys and interviews ............................................................................................. 233

2.3

Results .......................................................................................................................................... 233

2.3.1

Findings from Literature Review ................................................................................................ 233 2.3.1.1 IoT and WSNs .............................................................................................................................. 266 2.3.1.2 Survey findings and analysis ......................................................................................................... 28

2.4

Research Agenda .......................................................................................................................... 311

2.5

Conclusion ................................................................................................................................... 322 Declaration of generative AI and AI-assisted technologies in the writing process. ............................ 33

3.

A real-life optimization model for rose production planning with variable cycles and growing degree days curves: a case study in Ecuador .............................................................................................. 37

3.1

Introduction .................................................................................................................................... 39 3.2 Background on rose production ..................................................................................................... 41

3.3

Literature review of math models in crop planning and harvesting ............................................... 44

3.4

A Rose Production Model .............................................................................................................. 50

3.4.1

Notation ..................................................................................................................................... 50

3.4.2

Model Formulation .................................................................................................................. 522

| v

3.5

An illustrative real case ................................................................................................................ 533

3.6

Computational experience and analysis ....................................................................................... 544

3.6.1

Data collection ......................................................................................................................... 544

3.6.2

Model results ........................................................................................................................... 555

3.6.3

Sensitivity Analysis ................................................................................................................... 58

3.6.4

Managerial insights ................................................................................................................. 633

3.7

Conclusions .................................................................................................................................. 644

4.

Conclusions and Future Research ................................................................................................ 666

| vi

List of Tables

Chapter 1 Table 1.1. Outputs of this research…………………………………………………………………………………...11 Chapter 2 Table 2. Search protocol …………………………………………………………………………………………….22 Table 3. Literature Synthesis: applied technologies, main findings, and methodologies……………………….…...23 Chapter 3 Table 1. Agricultural Harvesting Models…………………………………………………….……………………....45 Table 2. Results Maria Bonita Farm-2021 (Business as usual) …...………………………………………………...55 Table 3. Results of the Maria Bonita Farm-2021 (Real versus model results) ....…….…….……………………….55 Table 4. Sensitivity Analysis: Influence of Minimum’s Variations………………………………………………….59 Table 5. Sensitivity Analysis: Influence of Production Cost………………….……………………….…………….60 Table 6. Sensitivity Analysis: Influence of Selling Price………….…...……………………………………………62

| vii

List of Figures

Chapter 1 Figure 1.1. Thesis Structure…………………………………………………………………………………………...10 Chapter 2 Fig.1 Main actors and stages for the flower business…………………………………………………………………19 Fig.2 SRL methodology stages………………………………………………………………………………………..21 Fig.3 Technologies Applied to Productivity in Flower Crops………………………………………………………....28 Fig.4 Structural summary of survey results……………………………………………………………………………30 Fig.5 Conceptual framework for the digital transformation of the floriculture activity……………………………...34 Chapter 3 Figure 1. Evolution of the Phenological Stages vs GDD. Figure used with permission of Falcon Farms S.A……….42 Figure

2.

(a)Timed

GDD

for Freedom rose.

(b)Screenshot of the

GDD

curve the freedom variety…………………………....................................................................................................................................43 Figure 3. Cutting curve (harvested stems) for Maria Farm…………………………………………………………….57 Figure 4. Production Cycle Timeline in Maria’s Farm………………………………………………..........................58 Figure 5. Financial Results versus Minimums’ changes……………………………………………………...………59 Figure 6. Overstock (dumps) and shortages vs minimums’ changes ……………………………………………. …...60 Figure 7. Gross profit vs production cost variations………………………………………………………………….61 Figure 8. Excess (dumps) and shortages vs production cost variations………………………………………………61 Figure 9. Gros profit vs production price changes…………………………………………………………………….62 Figure 10. Overstock (dumps) and shortages vs production price variations………………………………………...63

| viii

Abstract

Ecuadorian flowers, particularly roses, are globally recognized for their exceptional quality, representing approximately 11% of the global flower market valued at $8,5 billion. Within this sector, annual rose sales alone account for around $827 million. Despite its economic relevance, the floriculture industry faces increasing challenges related to climate change, phytosanitary regulations, production limitations, price volatility, and intense international competition. These pressures necessitate innovative strategies to enhance sustainability, efficiency, and profitability. This thesis, titled Industry 4.0 and Production Optimization in Ecuador’s Rose Export Sector: An Integrated Approach, seeks to offer the floriculture sector a new perspective—one that is oriented toward the strategic use of technological tools and process-oriented applications. The goal is to support producers in making better-informed decisions that contribute to the long-term sustainability and competitiveness of the business. The research highlights the urgent need for innovation to improve efficiency across the key stages of the flower production value chain. The study addresses two critical fronts: The optimization of rose production processes and the integration of Industry 4.0 technologies in flower farms, with a specific focus on the export business in developing countries.

First, the research explores the potential and challenges of applying Industry 4.0 technologies— such as smart agriculture, IoT, and data analytics—in the floriculture sector. Using a mixed-method approach that combines a systematic literature review and exploratory surveys conducted with flower farms in Ecuador and Colombia, the study assesses the impact of these technologies on productivity and sustainability throughout the entire value chain—from cultivation to post-harvest handling and distribution. Although adoption remains limited due to financial, technical, and human resource barriers, the findings reveal a substantial opportunity for implementation. Next, we developed a mixed-integer linear programming (MILP) model designed to optimize the rose-cutting process, incorporating growth curves, stem activations, production capacities, and business constraints. The model tested using real-world data from María Farm in Cayambe, Ecuador, proposes optimal cutting and harvesting schedules to maximize profit margins while ensuring timely demand fulfillment. Results demonstrate a significant improvement in business outcomes, including reduced waste (dumps) and increased stem sales, contributing to the longterm sustainability of rose production.

This research provides both theoretical and practical contributions by emphasizing the role of digital transformation as a strategic enabler and by presenting a concrete optimization model for rose production.

Keywords: Floriculture; Industry 4.0 Technologies, Smart Agriculture, Rose Production Model, Production Planning, Optimization, Mixed-Integer Programming Model.

Chapter 1

1. Preliminaries

1.1 Introduction

The global cut-flower industry represents a dynamic and highly competitive market, valued at over $8.5 billion in 2017, with projections estimating a compound annual growth rate (CAGR) of 4.2% to reach $53.9 billion by 2032. The Netherlands remains the global leader in flower exports, followed by Colombia and Ecuador, which contribute 15% and 10% respectively. In Ecuador, the flower industry—especially the production and export of roses—has become one of the country’s top five export sectors, generating more than 60,000 jobs and accounting for 10% of the national agricultural GDP. Roses alone represent 71% of flower exports, with the United States being the primary market.

Ecuador's prominence in the international flower market is largely due to its favorable climate and geographical conditions, which allow for year-round production of high-quality roses. However, the sector faces increasing challenges such as climate change, phytosanitary issues, price volatility, and heightened international competition. These factors place growing pressure on producers to improve operational efficiency and make accurate, timely decisions regarding the planning and execution of crop production, especially in processes such as flower cutting and harvesting. In this context, the adoption of digital tools and advanced technologies from the Fourth Industrial Revolution—such as the Internet of Things (IoT), artificial intelligence (AI), wireless sensor networks (WSN), and big data analytics—offers opportunities to transform the floriculture sector. These Industry 4.0 technologies can support growers across various stages of the value chain, from cultivation and harvesting to post-harvest handling and logistics, by enabling real-time monitoring, automation, and data-driven decision-making. While these tools have been extensively studied in traditional agriculture, their application in floriculture, especially in emerging economies like Ecuador, remains underexplored.

This research also aims to offer the floriculture sector a new perspective by highlighting the strategic importance of technological innovation and data-driven process applications. Specifically, it presents an integrated approach combining production optimization and digital transformation to support more efficient and sustainable decision-making.

1.2 Research Objective and Thesis Structure

This research aims to provide a road map to practitioners for the digital transformation in floriculture. The goal is to strengthen the sustainability and competitiveness of the business in the face of future challenges, enhancing decision-making and maximizing producers’ profitability. To do so, first, we critically review the technologies of the industry 4.0 applied in floriculture in emerging economies, identifying challenges and barriers for their implementation. Second, we provide a data-driven decision-making model that illustrates how the adoption of these models can result in significant productivity gains. Figure 1.1 illustrates the structure of this document.

Figure 1.1. Thesis Structure.

1.3 Structure and Organization of the thesis

This study comprises two complementary research streams. The first research stream, titled “The Role of Industry 4.0 Technologies in the Export Flower Industry: Insights from a Systematic Literature Review and Surveys in Emerging Economies,” adopts a qualitative methodological approach to analyze the adoption potential of Industry 4.0 technologies across the flower production value chain. Based on a systematic literature review and survey data collected from growers in Ecuador and Colombia, this stream identifies key adoption opportunities and constraints and proposes a conceptual business model in which Industry 4.0 technologies enable process integration, operational efficiency, and data-driven decision-making. This study is presented in chapter 2.

The second research stream, titled “A Real-Life Optimization Model for Rose Production Planning with Variable Cycles and Growing Degree Days Curves: A Case Study in Ecuador,” applies a

quantitative optimization approach through the development of a mixed-integer linear programming (MILP) model to optimize rose cutting and harvesting schedules. The model serves as a decision-support tool that can enhance harvesting efficiency, reduce waste, and improve the accuracy of harvested quantities. This study is presented in chapter 3. Chapter 4 provides a summary and consolidation of the key findings and insights from both research studies. It also outlines potential directions for future research in the field of supply chain management for flower growers in emerging countries.

1.4 Overview of publication outputs

The outcomes of the current research have been presented at three international conferences. One article has been published and officially presented at Expoflores (the trade association of the Ecuadorian floriculture industry). A second paper titled “A Real-Life Optimization Model for Rose Production Planning with Variable Cycles and Growing Degree Days Curves: A Case Study in Ecuador”, was submitted to the journal Smart Agricultural Technology on March 6, 2025, and is currently under review. The status indicates that the editor has advanced the manuscript to the peer review stage. Furthermore, three manuscripts were submitted to academic conferences during the preparation of this thesis. Table 1.1 below provides a summary of the publications and presentations derived from this research.

Table 1.1. Outputs of this research Reference Publication type Related chapter Mantilla, F., Mejía, G., & Tascón, D. (2025). The role of Industry 4.0 technologies in the export flower industry: Insights from a systematic literature review and surveys in emerging economies. International Journal of Results in Engineering, Article 104507. https://doi.org/10.1016/j.rineng.2025.104507 Research article Ch.2 Mantilla, F., Mejía, G., Suárez, N., Moreno, C., & Sarmiento, A. (2025). A real-life optimization model for rose production planning with variable cycles and growing degree days curves: A case study in Ecuador. Smart Agricultural Technology: https://track.authorhub.elsevier.com/?uuid=d03b0995-121f- 47e7-bfde-31954c2b8dba Research article (submitted- Under Review) Ch. 3 Mantilla, F., & Mejía, G. (2024). Smart agriculture for export flowers. Presented at the 12th International Conference on Production Research – ICPR Americas 2024. Paper recognized as the second-best submission at the conference.at https://link.springer.com/book/9783031777226 Conference ICPR 2024 (Indexed in the Book- Intelligent Production and Industry 5.0 with Human Touch, Resilience, and Circular Economy)

N/A

Reference Publication type Related chapter Mejía, G., Tascón, D. C., & Mantilla, F. (2023). Maximizing profits in the flower supply chain: The right mix of fixed and open market orders. In Proceedings of ICPR Américas 2022. ttps://link.springer.com/chapter/10.1007/978-3-031-36121- 0_26 Conference ICPR 2022 (Proceeding of the 11th International conference on Production Research Americas) N/A Mantilla, F., & Mejía, G. (2025). The Role of Industry 4.0 Technologies in the Export Flower Industry: Insights from a Systematic Literature Review and Surveys in Emerging Economies. Presented to the Board of Directors of Expoflores, Quito, Ecuador, March 28, 2025.

Conference at "Expoflores

– National Association of

Flower Exporters of Ecuador" N/A Mantilla, F., Páez, A., Suárez, C., Mejía, G., & Tascón, D. (2025). Comparison of traditional methods and machine learning for flower sales prediction. Paper to be presented at the 28th International Conference on Production Research (ICPR 28), Bogotá, Colombia, July 12–17, 2025. Conference ICPR 2028 N/A

Chapter 2

2. The Role of Industry 4.0 in the

Export Flower Industry

There is little doubt that Industry 4.0 and its applications will bring significant changes to the floriculture industry. While these technologies are still in the early stages of implementation in emerging countries, their impact is expected to become increasingly evident in the coming years, not only in business models but also in how work is performed, and operations are managed across various processes. These include crop infrastructure and the efficient use of key resources such as labor, which remains one of the primary cost drivers in this industry. This chapter1 also examines the influence of Industry 4.0 technologies on the productivity and sustainability of flower farms, covering the entire value chain from planting to post-harvest management and distribution. The aim of this research is to demonstrate to flower growers the potential impact of these technologies in optimizing the processes involved in flower production, harvesting, and export. This chapter also explores the opportunities and challenges associated with integrating Industry 4.0 technologies into floriculture, offering both theoretical and practical insights, along with recommendations to help industry stakeholders navigate this transition.

The content of this chapter is the paper submitted and published as: Fernando Mantilla – Gonzalo Mejía – Diana Tascón, (Mantilla et al., 2025a).“The Role of Industry 4.0 Technologies in the Export Flower Industry: Insights from a Systematic Literature Review and Surveys in Emerging Economies”, International Journal of Results in Engineering (RINENG)

2.1 Introduction

The cut-flower business is a significant global market, with $8.5 billion in sales recorded in 2017 (de Carvalho et al., 2022). The Netherlands leads the market, providing 43% of the global supply, while Colombia and Ecuador follow with contributions of 15% and 10%, respectively. These two Latin American countries have positioned themselves as key contributors in the international trade, with Colombia standing out as the thirdlargest exporter of flowers globally, closely followed by Ecuador. As the demand for fresh cut flowers grows, companies in these nations are progressively considering technological innovations to improve efficiency, sustainability, and competitiveness. Technologies linked to the fourth industrial revolution, such as Internet of Things (IoT), artificial intelligence (AI), Wireless Sensor Networks (WSN), robotics, and Blockchain, possess considerable potential to transform floriculture as they can be used in production processes, post-harvest management, and in general, the entire supply chains.

The adoption of Industry 4.0 technologies in agriculture is transforming traditional practices by introducing advanced tools and systems to monitor, automate, and optimize crop production. Technology advancements such as remote sensors, IoT, Wireless Sensor Networks (WSN), Bigdata, and robotics have been used to automate farms around the world (Adamides, 2020; Hassan et al., 2021). These technologies have leveraged the agriculture sector to evolve into a data-driven, intelligent, and autonomous connected system of systems (Lezoche et al., 2020). The application of Industry 4.0 technologies in agriculture has been widely studied, with focus on IoT, AI, and automation in both traditional farming and smart agriculture (Et-taibi et al., 2024; Pekkerieta et al., 2015; Van Hentenryck & Dalmeijer, 2024). For example, AI has been used to improve water management (Kamyab et al., 2023); Big Data supports agricultural practices by identifying patterns and optimizing resource use through IoT integration (Kamyab et al., 2023); WSNs combined with IoT platforms allow real-time data collection and improve risk management in smart agriculture (Et-taibi et al., 2024). Cloud platforms also play a role by improving data access and enabling advanced analytics for agricultural operations (Kamyab et al., 2023).

As mentioned above, Industry 4.0 technologies have been reported to improve resource management, increased efficiency, and reduced labor costs in many agricultural fields. However, studies focusing on the flower supply chain are limited (Heemskerk et al., 2020; Wani et al., 2023). Research specific to floriculture has a few authors have examined some aspects of automation (van Henten, 2006) and AI (Heemskerk et al., 2020) in greenhouse environments, but there is still a need to explore how these technologies can be applied across all stages of the flower supply chain.

This study investigates how these technologies can improve the performance of flower crops across various production stages, including post-harvest processes and export logistics. The research combines a systematic literature review with an analysis of exploratory surveys conducted among flower farmers in Colombia and Ecuador.

The organization of the paper is as follows: Section 0 presents an overview and theoretical foundation, exploring essential I4.0 technologies and their possible uses in the field of floriculture.

Background This section presents the essential components required to understand the convergence of floriculture and Industry 4.0 technologies. We start by examining the key variables to be regulated at greenhouse cultivation, including nutrients, water, temperature, pests, and light. Then, we describe the most commercially important flower species, highlighting their market characteristics followed by a description of the essential processes within the flower supply chain. This subsection covers aspects from planting through to post-harvest handling and distribution, emphasizing potential areas for technology-driven improvements. Last, we present a brief introduction to the technological innovations in agriculture, emphasizing tools and systems such as IoT, robotics, and AI that are transforming conventional farming methods.

2.1.1 Variables to control

In flower greenhouse cultivation, several variables are regulated to ensure growth, improve yields, and maintain plant health. These variables include nutrients, water and humidity levels, temperature, pest control, and light intensity, each playing a vital role in the success of the crop. Proper management of these variables helps not only in supporting plant development but also in mitigating risks such as disease or poor yield. Nutrients (Micro and Macronutrients). Plants require 17 essential nutrients for proper growth, most of which are derived from the soil, while a smaller proportion is obtained from the air (Hassan et al., 2021). The balanced supply of both micro and macronutrients is fundamental for supporting healthy crop development.

Water and humidity levels. Water aids in breaking down minerals, transporting nutrients, and enabling cell function. (Hatfield & Prueger, 2015). Both availability and adequate regulation of water are critical for sustaining plant growth and mitigating the negative effects of pesticides and fertilizers (Hatfield & Prueger, 2015). On the other hand, humidity is a variable that helps control the incidence of diseases and their effects on plants, which vary between productive and vegetative stages. This variable positively impacts plants during periods of high radiation by helping to mitigate the level of stress they experience.

Temperature. Temperature significantly affects the growth and development rate of plants. (Hatfield & Prueger, 2015) highlight that temperature control is essential, as it directly influences the pace of plant development and overall productivity. Pest Control. Management of pests, including insects, pathogens, and weeds is necessary to protect crops. The occurrence and impact of pests are closely related to weather and climate conditions, making pest monitoring and control vital to maintaining healthy plants (Juroszek & Von Tiedemann, 2013).

Light Intensity. Light is a critical variable in plant growth, as it provides the energy required by plants to regulate their development. Plants rely on light parameters such as intensity, composition, and direction to control their growth and adjust to environmental conditions (Fiorucci & Fankhauser, 2017).

Degree days. This is another critical variable in plant growth. It refers to the energy accumulated during the development process and is also known as physiological time or growing degree days (Monroy et al., 2014)

2.1.2 Perspectives in the global trade of flowers

Key players in the global market include European countries and the U.S. as major buyers, while the Netherlands, Ecuador, Colombia, Kenya, and Ethiopia are the leading producers and exporters. The top 5 leading companies in the cut flower market include The Queen’s Flowers Corp, Selecta Cut Flowers SAU, Sher Holland BV, Multiflora Corp, and Rosebud Ltd. (Grand View Research, 2024).

Floriculture crops include bedding plants, houseplants, flowering garden plants, potted plants, cut cultivated greens, and cut flowers. Typically, these flowers are used for decoration, aesthetics, and to exchange greetings. According to the Floriculture Market - Global Industry Analysis and Forecast (2024-2030) report, over 60% of all flowers and ornamental plants sold worldwide were used as gifts, 20% for weddings or funerals, and another 20% for home or office decor. The rose cut flowers segment led the market with the largest revenue share, accounting for approximately 32% in 2023 and lilies, renowned for their versatility and popularity, are among the most sought-after flowers worldwide (Market Analysis Report, 2023).

By distribution channel, the supermarkets and hypermarkets segment led the market with the largest revenue share, around 43% in 2023. According to the 2022 National Agricultural Statistics Service (NASS) report (Market Analysis Report, 2023), Americans collectively purchase approximately 10 million cut flowers daily, supporting a robust floriculture market valued at USD 6.43 billion across all retail channels, including supermarkets and hypermarkets (Market Analysis Report, 2023). The rise of online flower shopping reflects shifting consumer preferences towards digital transactions. The cultivation of ornamental flowers encompasses a wide variety of species, each presenting distinct growth needs. Main species in commercial floriculture include roses, chrysanthemums, carnations, gerberas, alstroemerias, and anthuriums (Dole & Wilkins, 2005). Roses (Rosa spp.) are the most extensively grown cut flowers worldwide, need meticulous management of temperature, humidity, and light conditions to achieve quality blooms (Reid & Jiang, 2012). The cultivation inside greenhouses facilitates control over various variables, especially temperature and light. Carnations (Dianthus caryophyllus), recognized for their resilience and diverse color palette, need precise nutrient management and pest control strategies. Automated irrigation and nutrient delivery systems are commonly used in this species(Jawaharlal et al., 2010). Chrysanthemums (Chrysanthemum spp.), popular as both cut flowers and potted plants, are photoperiodic species that require specific light durations to induce flowering. Growers can use smart lighting systems and blackout curtains, controlled by automated scheduling software to regulate blooming schedules and plant height (Dole & Wilkins,

2005). Chrysanthemums require careful management of climate conditions to ensure healthy growth and prevent disease. The control of humidity and temperature is particularly important to avoid common issues like botrytis and stemphylium. Implementing IoT-based climate regulation within greenhouses, such as dehumidification systems and environmental sensors, helps to maintain adequate conditions, required for both flower quality and longevity.

Alstroemerias (Alstroemeria spp.), or Peruvian lilies, are valued for their long vase life and colorful blooms. These plants grow in environments where root zones are kept cool and air temperatures are warm, making them suitable for greenhouse cultivation with automated temperature control systems (Dole & Wilkins, 2005).

Last, Anthuriums (Anthurium spp.) are tropical flowers known for their distinctive spathes. High humidity and warm temperatures are essential for their growth, and the use of controlled shading and air circulation systems is key for their cultivation (Higaki et al., 1984). The upcoming sections will explore in greater detail the practical applications of these technologies in addressing the specific cultivation requirements of these flower species.

2.1.3 Farm Processes

The growing of quality flowers includes several phases, from planting to distribution. This section provides a general description of these phases. Production process. The planting stage is the most critical part of the process and includes both the vegetative phase and the productive phase. The vegetative phase includes the growth stage, during which the vegetative system develops. In this phase, the plant develops on growth and energy accumulation, and it can take several weeks depending on the species to reach the productive stage, where the flower is harvested (Pérez et al., 2002).

Post-Harvest process. The process begins with the collection of cut flowers from the cultivation areas, followed by transport to the processing area, where they are received and classified according to their length, bud size, and quality. Once classified, the flowers are hydrated in tubs filled with water and preservative solutions to reduce stress and lengthen their vase life. After this process is complete, the flowers are transported to storage rooms. Among the operational decisions, harvesting, scheduling of production activities, storing, packing, and shipping are some of the most significant. Storage process. During the storage process, flowers are kept in controlled environments to maintain their freshness and quality (Pérez et al., 2002). This typically involves regulating temperature, humidity, and light conditions. The flowers are arranged in trays or boxes to prevent damage and ensure proper air circulation. Regular checks are conducted to monitor their condition. Proper storage is crucial to extending the shelf life of the flowers and minimizing waste before they are shipped for sale. Shipment process. The shipment process involves preparing the flowers for delivery to final customers and retailers. This includes packaging, labeling and all paperwork for exports. Before the shipment, orders are double-checked for quantities and documentation accuracy. The transportation means are chosen based on distance and destination to

preserve the quality of the flowers during transit. Normally, flowers are transported by plane, but in occasions ship transportation is also used (Mejía et al., 2022).

2.1.4 External Logistics Processes

The supply chain for the export of flowers is very complex as it involves several stakeholders, each playing a different role and often in conflicting situations. Flowers are handled by different actors (e.g. farms, transporters, logistic providers, airlines, police officers and, customs). A crucial element is the maintenance of the so-called cold chain which is essential to guarantee the flower quality throughout. The following are the main logistic processes.

Transportation to the airport of origin. This activity is normally carried out in airconditioned trucks. Trucks can either be owned by the flower company or hired. Temperature monitoring sensors or data loggers are placed inside the truck or sometimes in individual boxes. These sensors provide either real-time or recorded data, alerting handlers if there are important temperature deviations. In most cases, this occurs when truck drivers turn off the air conditioning system to save fuel (Mejía et al., 2019). Implementing integrated systems that account for the operational costs and constraints of both vendors and buyers can reduce inefficiencies and enhance overall flower supply chain performance (Sepehriar & Eslamipoor, 2024).

Transshipment. As said above, due to the high perishability of the flowers, the main transport is usually carried out by plane. This process occurs at the warehouses of logistic service providers who link customers, airlines and farms. In most cases, these logistic providers also perform additional tasks, such as quality control, handling, and packaging. They are also responsible for completing nationalization, clearing the shipment through customs, and delivering the flowers within the import country.

Delivery to the Distribution Channel. At this stage of the process, the product is delivered to the importer. Depending on the type of customer, deliveries may be made directly to a wholesale distributor or indirectly through an import trader. In the latter case, the product is delivered to the trader’s operational facilities for storage and subsequent distribution to retail clients, such as supermarkets, florists, and specialty stores. Most cut flowers have a lifespan of 7 to 12 days under moderate care. When the delivery is not made at the scheduled time, it shortens the flower's vase life and affects the value of the flower product (Tsai & Shen, 2024). The distribution decisions in the operational model include packing, storage and transportation decisions (Ahumada & Villalobos, 2011). Clearly, each of these stages plays a critical role in determining the quality and yield of the flowers produced, as shown in Fig. 1.

Fig. 1 Main actors and stages for the flower business

2.1.5 Technologies

This section aims to provide a concise summary of the key Industry 4.0 technologies and their potential to enhance the production and management of flower crops. In subsequent sections, we offer deeper context and concrete examples to show the impact of these technologies across various stages of flower production. Internet of Things (IoT). IoT refers to interconnected devices and sensors that collect and transmit real-time data from wireless sensors for efficient decision-making (Sekaran et al., 2020; Yang et al., 2021). In flower production, IoT can facilitate precision agriculture by enabling farmers to respond quickly to environmental changes, thereby improving crop health resource management, and risk mitigation (Et-taibi et al., 2024; Hosny et al., 2024; Mana et al., 2024; Morchid, Jebabra, et al., 2024; Morchid, Oughannou, et al., 2024).

Wireless Sensor Networks (WSNs). WSNs consist of spatially distributed autonomous sensors that monitor and record environmental conditions such as temperature, humidity, soil moisture, and light intensity, transmitting this data wirelessly to a central system for analysis and decision-making (Morchid, Oughannou, et al., 2024; Rodríguez et al., 2017; Rosero-Montalvo et al., 2020). WSNs provides real-time, continuous monitoring, which enables precise control over environmental variables that affect crop growth and health.

Blockchain Technology (BCT). The Blockchain platform provides a digital system and database to record the transactions along the supply chain (Park & Li, 2021). This decentralized database of transactions aims at transparency, reliability, traceability, and efficiency to the supply chain management (Park & Li, 2021). In floriculture, it can

also provide traceability, ensuring that information about flower origin, quality, and handling is recorded and shared (de Carvalho et al., 2022) .

Artificial Intelligence (AI) and Machine Learning (ML). Artificial Intelligence (AI) refers to systems or machines designed to simulate human intelligence, enabling them to perform tasks like recognizing patterns, solving problems, and making decisions (What Is Artificial Intelligence (AI)? | IBM, n.d.). Machine Learning (ML), a subset of AI, uses algorithms to analyze large datasets, identify patterns, and make predictions or decisions based on the data. AI and ML are increasingly used in agriculture to process large datasets collected from fields and greenhouses. AI and ML can assist in various aspects of flower cultivation, from predicting optimal planting times and disease outbreaks to automating nutrient delivery based on growth stages (Kamyab et al., 2023; Mana et al., 2024). A branch of AI also used in agriculture in species detection, counts and classification is Deep Learning (Estrada et al., 2024; Mann et al., 2022; Sun et al., 2021; X. (Annie) Wang et al., 2021) which is essentially an extension of traditional neural networks, incorporating advanced feature detection techniques. Robotics and Automation. These technologies encompass the use of automated machines for activities like planting, monitoring, and harvesting. In floriculture, robots can take on labor-intensive tasks, such as flower cutting and picking (Wani et al., 2023). Unmanned Aerial Vehicles (UAVs). UAVs, commonly known as drones, can be equipped with cameras and sensors, UAVs collect high-resolution imagery and environmental data, enabling farmers to assess crop conditions (Ahmed et al., 2023; T. Zhang et al., 2021). Aerial imageries are normally combined with ML and DL algorithms to extract information out of the images.

Cloud Computing (CC). Cloud Computing is a model that enables convenient and ondemand access to a shared pool of configurable computing resources, such as networks, servers, storage, applications, and services (Bumpus, 2013) . Cloud Computing facilitates the integration of IoT devices and machine learning systems, enabling real-time data analysis to optimize environmental conditions, improve traceability, and manage resources more efficiently (Et-taibi et al., 2024).

Big Data. Big Data technologies are used to analyze and derive insights, discover patterns from large datasets across various fields. In agriculture, these technologies have been used together with CC, ML and DL to store, retrieve and process large quantities of crop data and transform them into valuable information useful to farmers. Big data implies high speed capture, processing and analysis (S. Wang et al., 2023; R. Zhao et al.,

2018).

2.2

Methodology This study employs a mixed methodology: a Systematic Literature Review (SLR) alongside exploratory surveys and interviews carried out in Ecuador and Colombia farms. The literature review seeks to integrate current understanding regarding the application of Industry 4.0 technologies in flower crop production, whereas the

surveys offer perspectives from essential participants in the flower industry, particularly within emerging economies.

The literature review in this paper was carried out according to the methodology proposed by (Sauer & Seuring, 2023), following the stages presented in Fig. 2, and it is developed in Section 2.3. The exploratory surveys were carried out to evaluate the existing utilization and possible acceptance of Industry 4.0 technologies in the flower sector of Ecuador and Colombia. The selection of these countries is based on their important positions as flower exporters and their diverse strategies regarding technology adoption in the field of floriculture, the results of which are presented in section 2.2.2.

Fig. 2 SLR methodology stages. Based on (Sauer & Seuring, 2023).

2.2.1 Systematic Literature Review

Research questions definition. This study proposes the following research questions:

RQ1: What are the main applications of Industry 4.0 technologies in improving the performance of flower crops across different stages of production? RQ2: How have Industry 4.0 technologies, such as automation, artificial intelligence, and IoT, impacted the efficiency and sustainability of flower cultivation during the phases of cultivation, post-harvest, and distribution?

Defining the characteristics of primary studies, retrieving a sample of potentially relevant literature, and selecting the pertinent literature. To initiate the literature search for analysis, we formulated a search equation designed to support the research objectives. The main search equation incorporates keywords related to (i) Industry 4.0 and its associated technologies, and (ii) flower crops. The specific search equation is shown in

Table 2. SCOPUS was selected as the primary search database due to its extensive coverage of peer-reviewed literature across various disciplines. The study design resulted in the search protocol displayed in

Table 2.

Table 2. Search protocol.

Criteria Description Search equations TITLE-ABS-KEY ( ( "industry 4.0" OR "technology 4.0" OR "artificial intelligence" OR "Blockchain" OR "internet of things" OR "IoT" OR "big data" OR "data analytics" OR "cloud computing" OR drones OR "Unmanned Aerial Vehicle" OR "Unmanned Aircraft System" OR "Autonomous Aerial Vehicle" OR "Remote-Controlled Aircraft" OR robot* OR "digital twin" OR "digital simulation" OR "smart logistics" OR "smart manufacturing" OR "smart agriculture" OR "precision agriculture" ) AND ( floriculture OR "flower farming" OR "flower cultivation" OR "flower production" OR "floral agriculture" OR "ornamental horticulture" OR "roses" OR "orchids" OR "chrysanthemums" ) ) Type of files Journal Articles, Review Articles Database

SCOPUS

Exclusion criteria Studies that do not address the intersection of Industry 4.0 technologies and floriculture or ornamental horticulture. Documents that focus on non-technological agricultural practices or crops outside the floriculture sector. Studies that lack a clear methodological approach or insufficient information to answer research questions. Duplicate studies or articles with substantially similar content that are published in multiple sources.

Synthesizing the literature and reporting results. The data collected was organized based on defined criteria. Initially, the research was organized according to the specific Industry 4.0 technologies, including IoT, WSNs, ML, AI, robotics, and BCT. Each paper was subsequently assessed for its significance to floriculture, concentrating on its implications for flower cultivation, post-harvest processes, and supply chain management. Following the application of the exclusion criteria, the remaining literature was synthesized to identify prevalent themes, trends, and barriers for the implementation of Industry 4.0 technologies within the floriculture sector. This synthesis is presented in the Results, Section 2.3.

2.2.2 Exploratory surveys and interviews

The surveys were designed to gather different firsthand insights from flower producers on current practices, perceived benefits and challenges, and plans for adopting Industry 4.0 technologies. We used a closed-question format, primarily with multiple-choice options.

The survey structure and findings are presented in Section 2.3.1.2. These results also reinforce and add context to the theoretical insights gained from the literature review. Since the sample was small, we also interviewed technical managers of selected large farms about their perspectives and perceived barriers of Industry 4.0 technologies in floriculture. The interviews were semi-structured and conducted over the phone.

2.3 Results

2.3.1 Findings from Literature Review

Considering that our research is focused on flower crops, particularly those grown under greenhouses, most technological applications are used in production and harvesting processes, based on the understanding that these are the processes with the greatest opportunities to impact crop productivity.Technologies currently used in agricultural crops are widely applied across multiple stages of the supply chain, especially in production, harvesting, and logistics.

Table 3 synthesizes the studies clustered by technology type, highlighting each technology's applications and specific contributions to floriculture, and methodological approaches followed by the authors. We classified the papers according to three categories: Literature review/surveys/theoretical (LR/SU/OT), computational models (CM) and field tests (FT). The LR/SU/OT category self-explanatory. The CM category contains papers related to ML, DL or simulation models in which the focus is the development of the computational model. The FT category contains papers in which devices or technologies have been tested in the field. The devices or technologies can be either in a prototype stage or commercial applications.

Table 3. Literature Synthesis: applied technologies, main findings, and methodologies Technology References Applications Main findings Methodology IoT and Sensors (Atmaja et al., 2021; Bunpalwong et al., 2023; Burhan et al., 2023; Chang et al., 2021; Gagliardi et al., 2021; Hosny et al., 2024; Kamaruddin et al., 2019; Kumbi & Birje, 2022; Patel et al., 2023; Rosero- Montalvo et al., 2020; Salvini et al., 2022; Sangeetha & Periyathambi, 2024; Torres et al., 2023; Yang et al., 2021) Greenhouse monitoring, soil/crop monitoring, nutrient dosing IoT controls greenhouse environmental variables (Hosny et al., 2024).

IoT sensors assist in precise soil and crop monitoring (Gagliardi et al.,

2021).

IoT in cloud computing aids in virtual supply chain simulations for improved efficiency (Salvini et al., 2022).

Arduino-controlled wireless sensors enhance nutrient dosing in hydroponic flower systems (Sangeetha & Periyathambi, 2024) LR/SU/TH (Hosny et al., 2024) Implementation of sensors for monitoring humidity and temperature in realtime field experiments (Gagliardi et al., 2021) – FT Simulation (game), (Salvini et al., 2022) – CM Use of Arduino sensors to measure nutrients in hydroponic systems, (Sangeetha & Periyathambi, 2024) – FT Wireless Sensor Networks (WSNs) (Rodríguez et al., 2017; Rosero- Montalvo et al., 2020) Environmental monitoring, resource management WSNs enables precise climate control for rose cultivation (Rodríguez et al.,

2017)

WSNs facilitates greenhouse Experimental study of WSNs for precise climate control of roses using temperature and humidity sensors,

Technology References Applications Main findings Methodology monitoring, helping regulate growth cycles and resource efficiency (Rosero- Montalvo et al.,

2020)

(Rodríguez et al., 2017) – FT Experiments integrating WSNs in greenhouses to monitor growth cycles and optimize resource, (Rosero-Montalvo et al., 2020) – FT Blockchain (Chang et al., 2021; de Carvalho et al., 2022; Patel et al., 2023; Sumathi et al., 2022; Tsai & Shen, 2024; X. Zhao et al.,

2024)

Supply chain transparency, secure data management, flower classification, precision farming Blockchain improves flower classification accuracy by enhancing secure data transmission (X. Zhao et al.,

2024).

Combining IoT and Blockchain improves supply chain transparency, reducing labor costs (Tsai & Shen, 2024).

Blockchain enhances data integrity, helping provide crop recommendations based on real-time soil data (Patel et al.,

2023).

Proposal and implementation of an IoTand Blockchain-based model for enhancing the effectiveness of agricultural water management (Chang et al., 2021)

– FT

Development of a Blockchain system to classify flowers and analyze realtime data using encryption methods, (X. Zhao et al., 2024) – CM Artificial Intelligence (AI) (Fadzly et al., 2021; Herrera-Granda et al., 2020; Jha et al., 2019; Mana et al., 2024; Wehrens et al., 2024; Z. Zhang et al., 2024) Predictive analytics, disease detection, resource management AI and computer vision support sustainability by optimizing crop resource management (Mana et al., 2024).

Machine learning models predict disease outbreaks, reducing crop loss (Z. Zhang et al.,

2024)

Comprehensive review of AI applications, (Mana et al., 2024)

- LR/SU/TH

Analysis of disease outbreaks using machine learning models, (Z. Zhang et al., 2024) – CM.

Image Processing (Afonso et al., 2024; Caballero-Ramirez et al., 2023; Kwon & Heo, 2024b, 2024a; Li et al., 2024; Sivaraj et al., 2024; Sun et al.,

2021)

Flower classification and detection.

Defect detection.

Prediction of vase life.

Diverse ML and DL approaches for flower classification, phenological stage detection and defect detection improve on classical practices Classification and detection (Afonso et al., 2024; Li et al., 2024; Sivaraj et al., 2024; Sun et al.,

2021)

Defect detection (Caballero- Ramirez et al., 2023) – CM

Technology References Applications Main findings Methodology Vase life prediction (Kwon & Heo, 2024a, 2024b)

– CM

Cloud Computing / Big Data (Viet et al., 2021; R.

Zhao et al., 2018) Big Data platform for floriculture.

Supply chain and marketing.

Perspectives on the supply chain and the value of data in floriculture.

Need for protocols for R&D in flower industry LR/SU/TH studies Robotics and Automation (Pekkeriet & Van Henten, 2011; Priyadharshini et al., 2023; van Henten, 2006; van Henten et al., 2013; Wani et al.,

2023)

Labor automation, crop maintenance, harvesting Robotics is used for automating tasks such as harvesting in floriculture (Wani et al., 2023).

Robots streamline planting and maintenance in floriculture, cutting down on manual labor (Pekkeriet & Van Henten, 2011).

Development of an autonomous rover (Priyadharshini et al., 2023) – FT.

Unmanned Aerial Vehicles (UAVs) (Heemskerk et al., 2020; Patiluna et al., 2024; Rosero- Montalvo et al., 2020) Crop health monitoring, disease detection UAVs achieve up to

96%

accuracy in disease detection for orchids, outperforming manual inspections (Heemskerk et al.,

2020).

UAVs and

WSN

systems improve precision in monitoring greenhouse cycles, supporting floriculture management (Rosero-Montalvo et al., 2020).

Implementation of drones for disease detection in greenhouse crops, (Heemskerk et al., 2020) – FT.

Integration of UAVs and WSN to monitor greenhouse cycles and improve agricultural management (Rosero-Montalvo et al., 2020) FT.

UVAs and RFID method of inventory, experimental design (Patiluna et al., 2024)– FT.

We can observe that the most common technologies are IoT, WSN and AI (for flower classification). The following sections describe and discuss the use of these technologies in floriculture.

2.3.1.1 IoT and WSNs

IoT transmits information on the variables that affect crop production (Chang et al., 2021). This is done through remotely sensing main crop variables through remote cameras and photos (Ayaz et al., 2019). IoT is also used in prediction analysis, control of the use of pesticides, soil management, and conditions of crops (Atmaja et al., 2021).

WSNs are used in crop monitoring for early detection of unwanted crop states and offering support for onsite data collection (Ayaz et al., 2019). WSNs are used to measure variables of temperature, humidity, atmospheric pressure, and carbon dioxide (Rubanga et al., 2019). These sensors provide information for decision-making that affects crop yield and costs (Zervopoulos et al., 2020). They can be integrated with other tools and applications (Ayaz et al., 2019). Computational models, such as IoT applications for hydroponic nutrient monitoring (Sangeetha & Periyathambi, 2024) and Blockchain systems for classifying flowers and analyzing real-time data (X. Zhao et al., 2024), emphasize algorithmic advancements and data processing techniques. These technologies align well with the growing emphasis on sustainability in agriculture and floriculture (Eslamipoor & Sepehriyar, 2024).

Field test experiments, such as the use of WSNs for precise climate control in rose cultivation (Rodríguez et al., 2017) highlight their potential. The reviewed papers reported improved efficiency in the crops where these two technologies have been used. However, aspects such as costs, connectivity and technology adoption in emerging countries are missing in these reviewed papers, and a more holistic approach is needed before a full implementation in many crops.

Unmanned Aerial Vehicles (UAVs), commonly referred to as drones, have also emerged in the context of precision farming. Authors in (Heemskerk et al., 2020) explored the use of UAVs for greenhouse crop monitoring, specifically for orchid cultivation, where drones reported a remarkable 96% effectiveness in detecting plant diseases, greatly surpassing the accuracy of manual scouting. Their research highlights how UAVs can capture detailed aerial imagery, providing insights into plant health and enabling more targeted interventions to improve crop performance. Similarly, (Rosero-Montalvo et al., 2020) presented a monitoring system for environmental conditions in flower-cultivation greenhouses using UAVs and Wireless Sensor Networks (WSNs). This system significantly reduces data volume in the training set obtained by the WSNs while achieving high classification performance under real conditions, with accuracy rates of 90% and 97.5%, respectively. Additionally, (Patiluna et al., 2024) investigated the integration of Radio-Frequency Identification (RFID) with UAVs, conducting experimental designs to analyze how altitude and antenna power influence the accuracy of tag counts. As in the previous case of IoT and WSNs, outstanding performance has been reported but again, its wide world acceptance and adoption is still missing. Robotics has also impacted floriculture by automating processes. Authors (Pekkeriet & Van Henten, 2011) and (Van Henten, 2006) highlight the use of automation and mechanization in greenhouse environments, where the technology has reduced labor dependency and increased precision in crop management tasks like planting, watering, and climate control. The adoption of robotics is not limited to operational activities but also extends to quality control and disease detection. Authors such as (Dusadeerungsikul et al., 2019) and (Wani et al., 2023) discuss how robotic and cyber-physical systems can be employed to improve both the quality and quantity of flower crops. These technologies enable the automation of complex tasks such as pruning, flower recognition, and selective harvesting. In (Shree et al., 2019), the authors developed a robotic system for detecting and harvesting ripened marigold flowers, achieving an accuracy rate of 88%. Also, other autonomous systems (Kumar et al., 2016) use machine learning for plant recognition and

garden maintenance. In another application, (Asefpour Vakilian & Massah, 2017) present autonomous nitrogen fertilization systems that utilize machine vision-based analysis to optimize nutrient use in greenhouse crops. This precision management reduces overfertilization and minimizes environmental impact.

In terms of AI, several studies in floriculture describe the use of several DL algorithms for flower classification and defect detection in floriculture (e.g. (Afonso et al., 2024; Caballero-Ramirez et al., 2023; Kwon & Heo, 2024b; Li et al., 2024; Sun et al., 2021) However, most of them focus exclusively on the algorithms and use image databases. Some authors use samples from real crops (e.g. (Afonso et al., 2024; Caballero-Ramirez et al., 2023; Kwon & Heo, 2024a; Sun et al., 2021) ) but applications in the field are missing.

Fig. 3 presents a graphical description derived from the research included in the literature review of this document. It classifies the various technologies currently applied and highlights the stages and processes that have the greatest impact on productivity increases. Additionally, it illustrates their relationship with the key control parameters currently managed, which represent the main drivers or key variables affecting flower cultivation and management.

Fig. 3 Technologies Applied to Productivity in Flower Crops.

2.3.1.2 Survey findings and analysis

This section presents the results and analysis of a survey that we conducted to evaluate the current adoption, perceived benefits, challenges, and prospects of smart agriculture technologies in the floriculture sector in Ecuador and Colombia. The survey, comprising

34 questions, was organized around three primary areas of inquiry:

Current technology adoption: The first set of questions aimed to identify which Industry

4.0 technologies are presently utilized in flower production. This baseline information

provides a foundational understanding of how widespread these technologies are in the industry.

Perceived advantages and obstacles: To evaluate the impact of smart agriculture technologies, respondents were asked about the perceived benefits and challenges

associated with their use. Questions focused on production efficiency, crop yield, sustainability, and the barriers to broader implementation within their operations. Plans for future technology utilization: The final set of questions assessed the interest and readiness of farmers to integrate new technologies into their production methods. To conduct the surveys, 18 large companies in Ecuador and Colombia were contacted, all of them familiar with the use of 4.0 technologies. These companies currently cover 665 hectares of flower crops. We also contacted, mainly through Expoflores, small farms, but their managers refused to respond to the questionnaire as they said that they were not knowledgeable about the application of these technologies. The surveys were completed by technical managers and agronomists, experts in various products such as roses, carnations, pompons, and chrysanthemums. In addition to the surveys, a few technicians and managers were interviewed to either expand or clarify their responses and to provide additional insights. We describe our findings below:

The surveys showed that IoT and WSN were the most widely used technologies and these are already being used in production, post-harvest, storage, and transportation processes. IoT and WSN are also reported to help control and reduce the incidence of diseases and decrease flower waste. Most companies (56%) stated that these technologies also help improve crop productivity and quality and reduce costs. None of the surveyed 18 companies currently use Blockchain, drones, robots, image processing, machine learning, cloud computing or smart platform systems. About half of the respondents are unsure if or when they will adopt these technologies, while the other half foresee their implementation within a timeframe of one (1) to three (3) years. Even though technologies such as BCT, ML and robots are not being used in any of the

18 companies, over 80% of the respondents agree that these technologies may have a

significant impact on mainly reducing labor costs.

Most respondents (75%) strongly agree that both the shortage of skilled and specialized personnel are the major barriers to the implementation of technologies such as BCT, robots, Big Data, and Cloud Computing. When asked to elaborate on their responses, the interviewed managers (with technical background in agronomy or related fields) emphasized the need to create new job positions, and recruit personnel with technical expertise capable of developing applications and managing data in specialized ways. This includes proficiency in middleware development, designing custom Application Programming Interfaces (APIs) tailored to business needs, and knowledge of systems management. Another difficulty in implementation, identified by 78% of the interviewed personnel, is the low compatibility between systems (hardware/software), which increases investment costs and delays the adoption of the abovementioned technologies. Although to a lesser degree (50%), the responses also showed that high investment costs are also barriers that prevent the implementation of technologies. Surprisingly, resistance to change was not a barrier, at least for the educated professionals that we surveyed. Although most of the surveyed companies (72%) recognize the importance of using certain technologies, such as Big Data, BCT, WSN, drones, robotics, and image processing. However, these companies also identify connectivity as an issue and one of the most important barriers to implementing these applications, due to the low capacity of their systems to maintain stable connections with other networks and devices. The survey participants in Colombia and Ecuador claimed that, due to the location of some farms in isolated areas, there are difficulties with communication, coverage, and data transmission speed, which in turn affect the quality of real-time information. Although these barriers have been clearly identified, most agree on the need for significant

investments in the near future as part of the solution strategies. Other recommendations pointed out by the interviewees were:

• Connectivity and legacy systems are major barriers according to the manager of

a large farm. Although most farms have sensors and actuators, these are not part of integrated IoT or related systems. Implementing Industry 4.0 will require a full overhaul of all their legacy technologies.

• Most of the surveyed companies (75%) believe that significant advancements can

be achieved in monitoring crop conditions (Fertilization, Plant Health, Environmental Conditions) through these technologies.

• The majority of these companies (88%) recommend focusing efforts on

automating greenhouse management, emphasizing the automatic control and correction of environmental conditions (temperature, relative humidity, radiation, CO2 levels), which directly affect productivity. As one participant mentioned, "These systems may detect when temperature or CO2 levels are not optimal, but there’s very little we can do to optimize or correct them." One manager pointed out implementing these technologies as part of business strategies aimed at increasing competitiveness and an increase in the level of entry barriers. In the coming years, most of them anticipate a complete transformation in the industry, including production and new business models, which will require changes and redesigns in processes, structures, skillsets, workflows, and a new framework for contracts, supplier partnerships, and external specialized support as a condition for improving results In summary, the surveyed companies agreed that the present technologies improve crop productivity and have favorable impacts on reducing costs, flower waste, decreasing diseases, increasing efficiency and water use, and improving harvest date estimates. The surveys also revealed that most respondents agree that the impact of the applied technologies on social, environmental, and economic areas ranges from neutral to moderately positive and it will be a key factor for competitiveness, which will ensure the sustainability of the business.

In Fig. 4, we present a summary of the survey results, featuring a graphical description of the technological applications in flower cultivation. It highlights the stages with the highest usage, as well as the technologies that are not currently utilized.

Fig. 4 Structural summary of survey results

Summarizing, most I4.0 technologies used by companies in the selected sample and reported in the literature are limited to collecting and displaying information and on occasions exerting some form of control (e.g. a water pump may turn on to sprinkle flowers if the humidity goes below certain levels). This is the most common use of IoT and WSNs according to both our surveys and the literature review. Drones and cameras with image processing capabilities have also been reported in the literature and again their main use is to capture information that eventually will trigger human-driven actions. Robotics have also been used to automate some processes such as flower harvesting and classification. This is the replacement of human-intensive tasks. But drones and robotics have not been in use in the large farms that we contacted.

The use of predictive (i.e. forecasting) and prescriptive analytics was almost non-existent according to the literature and confirmed by our surveys. Big Data, ML (except for image processing) and formal optimization methods were rarely reported. Despite a wide variety of applications of Industry 4.0 technologies and their advantages, their use in these two countries (second and third in flower exports worldwide) is still in

its infancy. Only large companies have some knowledge about them and see potential in their adoption, and some have implemented data loggers and sensors connected through IoT devices to computing facilities; most smaller companies do not even know about these technologies and keep their traditional forecasting, cultivation and harvesting practices. Many companies are usually started by farm field workers (e.g. cutters, pickers, bunchers) who do not have the knowledge and/or the resources to implement these technologies. As the international markets start to demand not only cheaper and betterquality flowers but also more sustainable practices, the pressure builds up and may trigger some reactions, especially from the smaller companies.

2.4 Research Agenda

This research has laid a foundation for understanding the potential of Industry 4.0 technologies in floriculture; however, further exploration is necessary in specific areas. Future studies should focus on expanding the study's scope to include a broader range of flower production regions and larger sample sizes, which would provide a more comprehensive view of technology adoption in the flower sector. Several lines of research can be proposed based on the findings of the literature review and of the surveys. Some of them are:

Predictive analytics: IoT and WSNs have been used primarily for monitoring and control. However, their use in the flower sector can be more proactive and be combined with image processing, ML and automated decision making, especially in harvesting decisions. Cloud storage and computing can be used to run these models without onpremises investments in technology.

Prescriptive analytics: forecasting can be combined with optimization techniques to establish cutting and harvesting schedules. The idea is to maximize profits given not only crops but also market conditions (e.g. market prices, flower availability, etc.). Mathematical models that incorporate variables such as cutting times according to the flower growing times and other factors such as predicted degree days can significantly improve the operational profitability of the flower crops. These methodologies are absent from literature for the most part.

Supply chain operations: Blockchain has been hailed as a tool for transparency and traceability. However, many stakeholders in the flower industry may not be fully aware of Blockchain's potential benefits or lack understanding of how to implement it effectively. As we saw, some of the smaller farms do not even know about these technologies. Also, Blockchain technologies need huge investment in technology infrastructure and training. For many exporters, especially small and medium-sized companies, these costs can be prohibitive. Understanding the barriers needs additional research.

Human factors and sustainability. The new paradigm of Industry 5.0 has brought new challenges, especially the integration and adoption of technologies by human beings. In the flower industry, many issues are still unsolved: water management, for example, is a great concern. In the rose crop areas, the water consumption is drying out underground water sources; the use of pesticides and fertilizers are polluting water sources. Workers at the farms suffer from ergonomics related diseases. Also, migration to the cities and

urban centers is reducing the availability of qualified labor with the consequent risks to the flower farms. Research on I4.0 technologies can shed some light as to how to mitigate the negative impacts of the flower industry. Additionally, there is a need for comprehensive long-term evaluations of both environmental and financial impacts related to Industry 4.0 technology in the field of floriculture. further investigation is required to quantify these impacts and evaluate their long-term sustainability.

2.5 Conclusion

This paper conducted a thorough review of existing literature and analyzed survey results from farmers in Colombia and Ecuador to explore the role of Industry 4.0 technologies in the floriculture sector. The findings suggest that while the adoption of technologies such as IoT, artificial intelligence, robotics, and Blockchain is still evolving, there is significant potential for these innovations to transform supply chain management, postharvest handling, and floral production.

As part of the conclusions gathered, the main barriers to implementing these technologies were identified as high investment costs, the challenge of having specialized and qualified personnel, and the low compatibility of equipment (software/hardware) installed in different crop locations.

The surveys conducted helped prove that the largest companies in the sector or those with larger planting areas are currently leading the partial use of technological applications such as IoT, WSN, CC, and BC, with a stronger focus on production processes. The use of these applications contributes to improving crop performance, as they are employed in the control, monitoring, and traceability of variables.

Additionally, it was found that these variables are highly relevant in the stages of plant nutrition (water, fertilization), health (pest control), and growth (light, degree days). On the other hand, we also concluded that most small flower-producing companies in emerging countries like Colombia and Ecuador are not aware of the importance of these technologies, are far from implementing them, and, most concerningly, are unaware of the profound changes taking place or their impact on the business.

The research allowed us to infer that the floriculture sector still exhibits significant gaps in the implementation of fully connected and digitized agricultural operations through the use of Industry 4.0 technologies. This also highlights a tremendous opportunity to achieve greater efficiencies across all links in the flower production and commercialization business.

The results also indicate that merging IoT with artificial intelligence is particularly promising as it may result in advancements in environmental management, precision agriculture, and the automation of labor-intensive tasks such as harvesting. To sustain advantages in global markets, these technologies can help optimize resource utilization, enhance crop quality, and reduce operational expenses.

The study also indicates that the technologies currently used in the floriculture sector help producers manage information in an integrated manner, facilitating real-time decisionmaking, which results in increased production, reduced waste, and more efficient use of inputs such as fertilizers and chemical products. This efficient management also contributes to reducing the carbon footprint, optimizing raw material consumption, and ensuring the long-term sustainability of the business. Future challenges, according to the

information and testimonies gathered, should focus on the full application of existing technologies (e.g., robots, image processors) and new ones across all segments. This will require profound changes in the way crops are managed, executed, and controlled, leading to a significant transformation of business models.

The research conducted also helped us understand how the technological applications of Industry 4.0 can solve a wide range of relevant problems related to productivity and efficiency in the production, logistics, and marketing chains of flowers. Future research and studies related to the use of these technologies should focus on exploring new applications that help consolidate the integration of all the links involved in the business, from planting the plants to delivering the flowers to the final customer. Addressing these challenges through the various industry associations in the countries, with the support of universities and innovation centers, would allow the floriculture sector to fully harness the potential of Industry 4.0 and contribute to higher growth levels in the sector. The results also indicate that merging IoT with artificial intelligence is particularly promising as it may result in advancements in environmental management, precision agriculture, and the automation of labor-intensive tasks such as harvesting. To sustain advantages in global markets, these technologies can help optimize resource utilization, enhance crop quality, and reduce operational expenses.

The results also indicate that the advanced agricultural techniques in floriculture significantly improve efficiency, reduce waste, and promote sustainability, while also substantially increasing floricultural output. The broad adoption of technology faces numerous challenges, such as inadequate digital infrastructure in developing countries, the requirement for specialized expertise, and the substantial upfront costs associated with deploying new technologies. Addressing these challenges would enable the floriculture sector to fully harness the potential of Industry 4.0.

Declaration of generative AI and AI-assisted technologies in the writing process. During the preparation of this work the author(s) used ChatGPT in order to correct grammar and syntax. After using this tool/service, the author(s) reviewed and edited the content as needed and took(s) full responsibility for the content of the publication.

Conceptual Framework This section presents a conceptual framework aimed at conceptualizing the above drivers and barriers of the digital transformation of floriculture activity. This section is not part of the original paper. Figure 5 illustrates the proposed framework.

Fig. 5. Conceptual framework for the digital transformation of the floriculture activity. This conceptual framework illustrates the pathway through which the adoption of Industry 4.0 technologies can enhance the competitiveness of flower exporters. The model begins with the driving force of the analysis: the demands of the global market for high-quality, sustainably produced, and traceable flowers. These external conditions would motivate flower growers to adopt some of the Industry 4.0 technologies, the primary independent variables, including the Internet of Things (IoT), Artificial Intelligence (AI), Robotics, and Blockchain. However, according to our work, these technologies do not directly lead to competitiveness. Instead, their effect is directed through specific mediating variables, or technologies enablers. For instance, IoT facilitates real-time environmental monitoring, AI enables predictive analytics for harvesting schedules (as it will be seen in the next chapter), robotics automates manual tasks, and Blockchain provides supply chain traceability.

These mediating mechanisms, in turn, lead to tangible operational improvements, which are intermediate dependent variables. These technologies may control critical growing parameters like temperature and humidity. Predictive analytics helps establish cutting schedules, automation and traceability contribute to more efficient use of resources such as water and labor. The framework concludes that these combined improvements in operational performance—specifically in quality, efficiency, and sustainability—result in the final dependent variable: enhanced competitiveness in the global flower market. The model also incorporates the barriers identified in the research—such as high costs, a lack of skilled personnel, and poor connectivity—which negatively moderate the entire process by hindering both the initial adoption of technologies and the effective operation

of the mediating mechanisms. Based on the above framework we propose the following guidelines.

1.

Strategic alignment and development of organizational capabilities The adoption of Industry 4.0 technologies should be conceived as part of a comprehensive business strategy aimed at improving competitiveness. This transformation must be accompanied by organizational changes, including the redesign of processes, structures, workflows, and collaboration models with suppliers and technology partners. In this context, it is essential to involve personnel with adequate training in these technologies from the early stages of implementation. In essence, farms must determine which are the strategic goals they want to achieve and what are the barriers and challenges for these implementations(El-Telbany et al., 2020).

2.

Preliminary assessment of the technological maturity of the farm (starting point) Before initiating any process related to the implementation of Industry 4.0 technologies, the farm must conduct a comprehensive assessment of the current state of the farm in order to determine its level of technological maturity with the TOE (Technology- Organization-Environment) framework(Senna et al., 2023; Tornatzky & Fleischer, 1990). The findings of this study indicate that, although many farms are equipped with sensors and actuators, these technologies often operate in isolation rather than as integrated systems. Consequently, any digitalization strategy should begin with a realistic evaluation of legacy systems and the definition of a phased modernization roadmap, thereby avoiding fragmented or low-impact implementations. Such an assessment should identify existing gaps, particularly in critical areas such as connectivity, technological infrastructure, available equipment, control architectures, and the level of integration among current systems.

3.

Phased implementations Based on our experience and the findings of the study, farms should move from basic automation and data capture toward a fully automated and integrated data-driven decision-making architecture that responds in real time to the stringent market demands. The simplest implementations are automated greenhouse management systems capable of detecting and correcting deviations in parameters such as temperature, relative humidity, radiation, and CO₂ levels through closed-loop control mechanisms (descriptive analytics). At more advanced stages, technologies and methods for predictive analytics for forecasting both market and farm conditions are obvious candidates. Tools designed for labor allocation, cutting plan definition, harvest date determination, and production capacity estimation integrating agronomic, operational, and market variables help reduce mismatches between supply and demand, minimize waste, and enhance customer service levels (e.g. (Hartmann et al., 2017).

4. Technological integration and progressive replacement of legacy systems

At an advanced level of digital maturity, priority should be given to integrating existing systems into unified data management and analytics platforms. The continued reliance on non-interoperable legacy systems remains one of the main barriers to the effective

adoption of Industry 4.0 technologies(P. Senna et al., 2023). Accordingly, modernization strategies should promote the progressive upgrading or replacement of legacy systems, ensuring interoperability among sensors, actuators, analytics platforms, and management systems. This integration is essential to maximize the value of generated data and to support both operational and strategic decision-making.

Closing Remarks on the Recommendations In conclusion, the recommendations presented in this section provide a structured and progressive roadmap for the adoption of Industry 4.0 technologies in floriculture within emerging economies. The proposed sequence from initial diagnosis to full technological integration addresses both technical and organizational dimensions, recognizing that digital transformation is not solely a technological challenge but a strategic and managerial process.

By prioritizing diagnostic assessment, capability development, and early interventions with direct impact on flower production, producers can reduce implementation risks and ensure that technological investments generate tangible productivity gains. As technological maturity increases, the integration of automation, data analytics, and decision-support systems enables more accurate forecasting, improved resource allocation, and enhanced operational efficiency.

Ultimately, the full integration of systems and processes constitutes the foundation for a data-driven, resilient, and competitive floriculture model. These recommendations reinforce the study’s hypotheses by demonstrating that the effective adoption of Industry 4.0 technologies depends on a phased approach aligned with organizational readiness, production priorities, and strategic objectives(Liu et al., 2021; Roumeliotis et al., 2024). This framework offers both academic and practical value, serving as a reference for researchers, practitioners, and decision-makers seeking to advance digital transformation in the floriculture sector.

Note on the study:

Although our research allows for cross-referencing field results and validating them through an exhaustive review of the literature on the application of technologies, it is clear that the sample was small. In future works it will be crucial to include small-scale growers even if they lack knowledge of I4.0 technologies. By doing so, the study will have greater validity, reliability, transferability, and generalization of the results.

Chapter 3

3. A real-life optimization model for

rose production planning with variable cycles and growing degree days curves: a case study in Ecuador The rose production process is a vital component of the floriculture industry, with a significant impact on the global economy and flower trade. This research presents a Mixed-Integer Linear Programming (MILP) model specifically for rose crops, incorporating elements such as growth curves, stem activation patterns, crop capacity, and business constraints. The model generates optimal cutting schedules and harvesting dates to maximize profits while ensuring timely fulfillment of demand. To the best of our knowledge, models addressing cutting schedules for rose production have not been previously proposed in the literature. The actual results of the model demonstrate a significant increase in business margins, achieved by reducing waste (dumps) and increasing the number of stems sold. Additionally, this chapter2 provides insights for stakeholders and business leaders, illustrating how a tool like the MILP model can be applied to operational processes to drive profitability and support more strategic decision-making. From a digital transformation perspective, this model represents a shift from empirical and reactive production planning toward a data-driven, algorithmic

2 The content of this chapter is the paper submitted on November 2025, currently under review as: Gonzalo Mejía, Fernando Mantilla, Natalia Suárez, Carlos Moreno, Alfonso Sarmiento, “A Real-Life Optimization Model for Rose Production Planning with Variable Cycles and Growing Degree Days Curves: A Case Study in Ecuador.”, at Journal:

Smart Agricultural Technology:

https://track.authorhub.elsevier.com/?uuid=d03b0995-121f-47e7-bfde-31954c2b8dba

decision-making framework. In this sense, the MILP formulation can be the first layer of a digital twin for rose production, where real-time data from greenhouses (e.g., IoT-based climate monitoring) and market information can feed optimization engines to continuously update cutting schedules and profitability projections. Thus, the chapter not only presents a mathematical model but also illustrates how digital transformation can improve processes in floriculture through prescriptive analytics and integrated decision platforms.

3.1 Introduction

The market size of the global cut flowers industry is predicted to rise by a compound annual growth rate (CAGR) of 4.2% to USD 53.9 billion in 2032 (Spherical Insights LLP, 2023). The export flower sector is considered an important currency generator for both developed and emerging economies (Cedillo Villavicencio et al., 2021). In emerging countries, and particularly in Ecuador, the flower industry ranks among the top four export sectors, according to the report by Expoflores, the National Association of Flower Producers and Exporters of Ecuador (Expoflores, 2024). The Expoflores report indicates that Ecuador accounts for 9% of the global value of flower exports, following the Netherlands and Colombia. Ecuadorian roses represent 76% of exported flowers, with the United States as the main destination, accounting for 29% of total exports. It is worth noting that roses are the most in demand floral product in the U.S., where annual imports reached approximately USD 222 million in 2024. Rose production in Ecuador exceeds 5,000 hectares distributed across several regions, mainly in the Andean province of Pichincha. The industry comprises around 1,000 flower exporters, of which 86% export less than USD 1 million annually. The industry contributes approximately 11% to the country’s agricultural gross domestic product (GDP). (Expoflores, 2024). Roses from Ecuador are among the best in terms of quality, thanks to ideal conditions that include year-round favorable temperatures and daily sunlight exposure of about 12 hours (Villalobos et al., 2012). However, internal and external competition has placed strong pressure on flower growers to improve efficiencies and meet market demands (Guaita-Pradas et al., 2023; Navarrete et al., 2018). There are also several challenges that hinder their competitiveness (Vanegas López et al., 2017): biological and phytosanitary issues increase the need for agrochemical applications (Faust et al., 2016; Vasco et al., 2025). The industry remains highly labor-intensive, facing rising wage costs and seasonal shortages of skilled workers, particularly during peak periods such as Valentine’s and Mother’s Day (Faust & Dole, 2021). Moreover, deficiencies in postharvest handling, cold-chain logistics, and resource management (water, fertilizers, and energy) also constrain overall efficiency (Babalola et al., 2011). Finally, operational inefficiencies also persist, as many farms rely on empirical production planning rather than data-driven planning tools, resulting in production imbalances, waste, and unmet demand (Drechsler & Holzapfel, 2022; Mantilla et al., 2025b). Addressing these challenges requires the integration of advanced technologies and optimization models to enhance productivity and ensure long-term sustainability in the Ecuadorian floriculture sector (Mantilla et al., 2025b).

Regarding production planning, one of the most important decisions that farms must make is to determine a cutting schedule, that is, to define the number of stems to be harvested during the fifty-two (52) weeks of the year. This is a challenging task because the cutting schedule must respect the biological flowering cycles while simultaneously meeting a highly fluctuating demand with pronounced peaks on Saint Valentine’s Day and Mother’s Day. Unlike other agricultural products, in which harvests are determined by the season, the flowering cycles of roses depend on previous cuts. Every time a stem is cut, whether the flower is in bloom or not, a new cycle begins. Since cutting schedules are typically established empirically and without analytical support, this often results in considerable quantities of both excess and shortage of flowers. Excess flowers— commonly referred to by farms as “dumps”—are those that cannot be exported due to insufficient market demand and are systematically discarded, generating no revenue and incurring additional operational costs. In contrast, shortages lead to backordered customer

orders, forcing farms to procure the required quantities from third parties at substantially higher prices.

As such, to meet demand throughout the year, the number of stems to be cut must align with customer orders. To achieve this, the farm must develop a production plan composed of two phases: planning and cut scheduling. The planning phase begins with the requirements provided by the sales department, which specifies the quantities of flowers required for the upcoming year. These quantities constitute the farm’s “production budget.” In the cut scheduling phase, this gross annual estimate is translated into a weekly cutting plan. However, this is easier said than done, as the peak periods of Saint Valentine’s Day in February and Mother’s Day (separated by a flowering cycle of approximately 88 days) make scheduling particularly challenging. If the farm decides to harvest all its flowers for Valentine’s Day—taking advantage of both high demand and higher prices—additional peaks will occur in May, August, and November, leading to shortages in the remaining months of the year. This situation is undesirable for two reasons: first, farms have binding contracts with wholesalers and retailers throughout the year to supply negotiated minimum quantities (this is called “every-day” sales); and second, demand during August and November is typically low and this fact leading to significant quantities of dumps. Therefore, farms must strike a balance between concentrating production during high-demand seasons (February and May), which may result in excess and dumped quantities during subsequent peaks, and maintaining sufficient flower availability to fulfill contracts during the low season, albeit at lower prices. Moreover, roses can only be stored for short periods, making inventory carrying unfeasible. Unlike manufactured products, even with accurate sales forecasts, it is virtually impossible to avoid either excess or shortage of flowers at certain times of the year. As said in the abstract, dumps and shortages represent expenses of hundreds of thousands of US dollars per year in a single farm alone. According to senior managers in the industry, devising the “optimal” cutting schedules is one of the “holy grails” in floriculture.

An advantage of roses and other perennial crops, unlike seasonal agricultural, the production cycles of roses can be adjusted. The normal 88-day production cycle may be shortened or even interrupted (by cutting stems) to accommodate demand or to avoid dumps. One disadvantage is that the number of combinations of cutting schedules can be huge. This highly combinatorial problem is a special case of production planning and scheduling with added complexities that will be described below. This fact fully justifies the use of optimization tools for this problem.

To the best no previous work has developed optimization models for flower cutting schedules. There are hundreds of models in the literature that develop planting and harvesting schedules for a wide variety of agricultural products but not for the specific case of roses which have unique market, growth, flowering and harvesting characteristics. This research aims to answer the following research question (RQs): RQ1: What is the cutting schedule throughout the 52 weeks of the year that maximizes profits?

RQ2: What are the desired minimum selling quantities that the farm should commit and negotiate with its customers in the low season?

RQ3: What are quantities of dumps and backorders resulting from the optimization model in the year?

To address these questions, our research proposes a mixed-integer linear programming (MILP) model for the production planning of the “Freedom” variety rose. The novelty of our proposed model lies in the fact our model uses the growing constraints of perennial crops, variable production cycles and growing degree-day curves to estimate quantities and harvest dates. Our model also considers constraints such as meeting weekly demand, overall farm capacity, and other related to business operations such as backorders and minimum quantities. This paper extends beyond a purely academic exercise as the farm where it was tested is seriously considering its full implementation. The findings of this model are scheduled to be presented by the authors at an upcoming panel discussion during a prominent floriculture conference in the upcoming months. This paper was organized as follows: Section 2 provides a background of the flower production process. Section 3 presents a review of the literature on scheduling and planning for flower and other agricultural products to examine similarities, differences, and research gaps. Section 4 introduces the scheduling problem and presents the proposed model in detail. Section 5 illustrates a real case of a rose farm located in Ecuador. Section

6 presents computational tests and discussion of the results. Finally, conclusions and

future works follow in section 7.

3.2 Background on rose production

The rose production process consists of two (2) main phases: phase 1, also known as the “vegetative” phase, includes activities that span from planting to the first pinch or initial pruning, and includes stem growth and bending. This first phase may take between 20 and 24 weeks, depending on the variety. Once this phase is completed, phase 2, also known as the “productive” phase, begins. The duration is determined by energy accumulation. A production cycle is defined as the period between two successive rose cuts, normally between two rose blooms. In the remainder of the paper, the term “cycle” denotes the “production cycle”.

The duration of the production cycle depends on Growing Degree Days (GDDs) rather than time. GDDs are a measure of the accumulation of energy necessary for estimating plant growth stages. GDDs are calculated by subtracting a base temperature from the average of the daily maximum and minimum air temperatures, expressed in °F or °C. The base temperature represents the threshold below which the flowers do not grow. During the production cycle, roses go through several phenological stages. Figure 2 represents the average GDDs that a rose plant requires to reach a phenological stage from the last cut. For example, 430 GDDs are required on average to reach the “rice” stage, 610 GDDs are required for the “ball” stage and 780 GDDs for the cut-off point. The GDD constraint makes it possible to link the growing conditions with the model’s decision variables, ensuring that the production planning reflects the actual behavior of the crop

Figure 2. Evolution of the Phenological Stages vs. GDD. Figure used with permission of María Farm.

However, flowers may bloom before or after the theoretical pre-calculated required GDDs. This fact leads to a probability density-like curve in which a data point corresponds to the probability of bloom vs. the GDDs. For our mathematical model, the original GDD curve was transformed into a timed-GDD curve from the expected values of cumulative GDD throughout the production cycle. The values in the x axis were calculated as 𝑐𝑢𝑚 𝐺𝐷𝐷 𝐺𝐷𝐷 𝑑𝑎𝑦̅̅̅̅̅̅ where 𝑐𝑢𝑚 𝐺𝐷𝐷 are the cumulative growing degree days and 𝐺𝐷𝐷̅̅̅̅̅̅ 𝑑𝑎𝑦 are the average GDDs absorbed per day. In the María Farm, this average is 8.7 GDDs/day. Said that, the goal of 780 GDDs required to harvest the Freedom rose (see Figure 2) is achieved 88 days on average after the last cut at this farm. An example of the timed-GDD curve provided by María Farm is shown in Figure 3a. The 𝑥-axis represents the average number of days since the last cut, while the 𝑦-axis shows the values of the probability density-like function. The area under the curve between two given days represents the probability that a rose is in bloom and ready for cutting. Figure 3b is a screenshot of the GDD curve established at the farm.

cm

(a)

(b) Figure 3. (a) Timed GDD for the Freedom rose. (b). Screenshot of the GDD curve of the Freedom variety.

The cumulative expected number of flowers harvested up to week 𝜏 depends on quantities cut at the start of week 𝑠 and on the GDD curve. This expected value can be computed as follows:

𝜁(𝑠, 𝜏) = ∫𝑄𝑠,𝑡 ψ𝑡−𝑠𝑑𝑡 𝜏 𝑠

Eq (1) Where 𝑄𝑠,𝑡 is the quantity of stems cut in week 𝑡 and whose cycle started in week 𝑠 and ψ𝑡−𝑠 is the GDD probability density function at time 𝑡−𝑠. The cumulative total quantity of flowers 𝐻𝑐𝜏 harvested up to week 𝜏 is given by:

𝐻𝑐𝜏= ∑𝜁(𝑠, 𝜏 𝑠 𝜏) Eq (2) The total quantity of flowers harvested in week 𝜏 will be given by:

𝐻𝜏= 𝐻𝑐𝜏−𝐻𝑐𝜏−1 Eq (3) If we discretize the GDD curve, the expression for the total harvested quantity of flowers harvested in week 𝜏 will be:

𝐻𝜏= ∑𝑄𝑠,𝜏 ψ𝜏−𝑠+1 𝑠

Eq (4) This timed GDD curve is essential for the development of our model.

3.3

Literature review of math models in crop planning and harvesting Although the world export flower industry represents billions of dollars per year, the number of studies aimed at developing models to optimize production plans in floriculture is relatively few and outdated. The literature, however, is extensive in crop planning and harvesting of agricultural products as shown in several reviews, e.g. (Ahumada & Villalobos, 2009; Kusumastuti et al., 2016; Taşkıner & Bilgen, 2021). The crop planning problem consists of determining what crops to plant, when, and with which management practices to maximize expected long-term profit while accounting for risks, soil dynamics, and other uncertain factors such as weather and prices; the harvesting problems consists of determining an optimal schedule to collect the yield in a time window that is constrained by resources such a workers and machinery. In general, the decision variables of the above models are related to the harvested quantities, timing of the harvesting and the resource usage. Normally, the harvest is derived from the crop plan. This paper focuses on the harvesting (cutting in this case) aspect. Table 1 shows a summary of the findings.

From the table, we can observe that there are very few papers that use optimization models to establish planting and harvesting schedules in the flower industry. Among the few, Caixeta-Filho et al. (2000) proposed a model for optimizing gladiolus bulb production using a linear programming model to maximize economic returns. Later, Caixeta-Filho et al. (2002) describe the implementation of a decision-support system (DSS) based on a linear programming model for optimizing production and trade in the lily flower business. Both papers were implemented in Brazilian companies. More recently, Hoogeveen et al. (2019) proposed a model that establishes the number of mother plants to meet a given demand of cuttings per week. The only paper on rose production with mathematical models was the one of Ruiz-Torres et al. (2012) who presented a math model to establish rose planting (not cutting) schedules. This is the number of rose plants to sow and remove in a given period over a multi-year planning horizon. The objective function was maximizing profits.

In other agricultural products, most of the reviewed papers study the harvesting problem only but others combine crop planning with harvesting. Products include perishable fruits such as tomatoes (Rocco & Morabito, 2016), wine grapes (Ferrer et al., 2008; Varas et al., 2020), bananas (Caicedo Solano et al., 2020), pommes (Gómez-Lagos et al., 2021; Montenegro-Dos Santos et al., 2023), and blueberries (Fathollahi-Fard et al., 2023). Other harvesting models investigate non-perishable products such as sugar cane (Aliano Filho, et al., 2023; Jarumaneeroj et al., 2021), olive oil (Herrera-Cáceres et al., 2017), corn (Chen & Ryan, 2023; Custodio et al., 2024; Khalilzadeh & Wang, 2022; Y. Zhang & Swaminathan, 2020), guayule (Mahdavimanshadi et al., 2024; Yao et al., 2023; Zuniga Vazquez et al., 2021) and palm oil (Escallón-Barrios et al., 2022).

Table 4. Agricultural Planning and Harvesting Models Authors Description Objective Function Category Product /species Model Flowers and other perennial crops

(Caixeta-Filho et al.,

2000)

Determines the quantities to plant of gladiolus bulbs subject to growth cycles, bulb availability and space limitations.

Economic Crop planning Gladiolus LP (Caixeta-Filho et al.,

2002)

Describes the implementation of a decision-support system to optimize the production planning and sales strategy for a commercial lily flower farm Economic Crop planning and harvesting Multiple lily varieties

MILP

(Ruiz-Torres et al., 2012) Develops tactical production planning models to maximize profit by allocating land and workforce for different rose varieties over a multi-period horizon.

Economic Crop planning Roses

MILP

(Hoogeveen et al., 2019) Determines the minimum number of mother plants and the harvest schedules required to meet a forecasted demand for cuttings.

Biological Crop planning and harvesting Ornamental plants (Dümmen orange) LP (Zuniga Vazquez et al.,

2021)

Develops a model for multi-crop production planning and machinery scheduling, accounting for crop rotation, irrigation constraints, and machinery routing.

Maximizes the farmers’ net present value (NPV) Crop planning and harvesting Guayule and Guar

MILP

(Yao et al., 2023) Determines the harvest quantities and the machinery schedules under time windows, resource constraints, machinery routing, and demand. Minimizes total harvest costs Harvesting, routing Guayule

MILP

(Mahdavimanshadi et al.,

2024)

Calculates guayule harvest plans and machinery schedules under severe drought scenarios Minimize total expected costs Crop planning and harvesting, routing Guayule, cotton, wheat

MILP - SP

Other agricultural products

(Widodo et al., 2006) Develops an analytical model of the periodic flowering and harvesting processes of perishable agricultural products Minimize deviations of demand vs production Harvesting Generic for fresh agricultural products.

LP (Ferrer et al., 2008) Wine grapes Minimize cost Harvesting, labor allocation and routing Wine grapes

MILP

(Nagasawa et al., 2009) Formulates a cooperative harvest-delivery with coordination of harvesting schedules under perishability, delivery lead times, and flowering constraints.

Maximize the constant daily consumption level across all markets Harvesting Generic perishable products

MILP

(Rocco & Morabito,

2016)

Determines the varieties, quantities, cultivation and harvesting schedules, transportation modes and processing units in tomato production and processing.

Minimize total production and logistics cost Cultivation and harvesting.

Tomato LP

(Herrera-Cáceres et al.,

2017)

Defines a harvest schedule of land zones under resource and technical constraints Maximize the extraction of olive oil in a mill.

Harvesting Olives

MILP

(Boussios, Preckel, Yigezu, Dixit, Akroush, M’hamed, Annabi, Aw‐ Hassan, et al., 2019) Incorporates farmers' adaptive responses to weather events to optimize long-term crop management and rotation plans.

Maximize the expected discounted sum of current and future profits Crop planning Generic

SDP

(Varas et al., 2020) Proposes a model for wine grape harvest scheduling with multiple decision makers Minimize operational costs and maximize grape quality Harvesting Grape Multi-objective MILP (Y.

Zhang & Swaminathan, 2020) Develops optimization models to schedule the planting and harvesting of corn hybrids with different constraints Maximize expected total profit and total costs Crop planning and harvesting Corn hybrids

SDP

(Caicedo Solano et al.,

2020)

Solves planning of agricultural production systems in the stages of sowing, crop maintenance and harvesting.

Minimize total cost and waste Crop planning and harvesting Banana crops

MINLP

(Gómez-Lagos et al.,

2021)

Defines the harvest schedule that in multiple fruit orchards that share labor and other resources. Minimizes total costs unripe fruit loss, and total harvest duration.

Harvesting Multiple fruit orchards GRASP Metaheuristic

MILP

(Khalilzadeh & Wang,

2022)

Develops optimization and machine learning models to schedule the planting and harvesting of diverse corn hybrids for two storage capacity cases, accounting for uncertainty in GDD curves. Minimize the sum of absolute differences between the weekly harvest quantity and the storage capacity Crop planning and harvesting Corn hybrids MILP and heuristics (Escallón-Barrios et al.,

2022)

Proposes a set of interconnected models aimed at strategic (harvest cycles), tactical and operational decisions.

Minimize costs Harvesting Oil palm MILP-Simulation- Data analytics (Fathollahi-Fard et al.,

2023)

Proposes a framework that balances cost, fruit quality, and schedule compactness for harvesting for real-world decision-making for large fruit exporters Optimize multiple objectives: maximize profit, minimize waste and carbonemission.

Harvesting Blueberry.

Multi-objective fuzzy mathematical model and a new metaheuristic algorithm variant of

NSGA-II

(Aliano Filho, et al.,

2023)

Models for harvesting sugarcane under different constraints and scenarios.

Minimize grinding losses, unharvested cane, and equipment movement costs, Harvesting Sugar cane

MILP

(Montenegro-Dos Santos et al., 2023) Introduces a non-myopic rolling horizon scheduling and a rescheduling methodology for olive trees’ harvests Minimize schedule variability and contractual disruptions.

Maximize production.

Harvesting Olive trees

MILP

(Albornoz & Vera, 2023) Defines management zones for harvest planning, and coordination between producers and wholesalers.

Minimize total cost Harvesting and land allocation Hierarchical supply chain SP bilevel (López et al., 2024) Proposes supply chain model for smallholder farmer cooperatives to optimize crop planning over Maximize profit and food security Crop planning and harvesting Vegetables and grains MILP and SP

multiple seasons. Decision variables are related to planting and harvesting quantities. (Custodio et al., 2024) Presents a model for harvest scheduling and land allocations for multiple crops. Maximize regional agricultural profitability Harvesting and land allocation Maize and rice LP (Chan et al., 2025) Proposes model to design a circular (closed-loop) supply chain for perishable agricultural products, under perishability and recovery (by-product) flows.

Maximize profitability, minimize waste, maximize average freshness / quality Production, harvesting and distribution.

Generic

MINLP

Conventions:

MILP: Mixed Integer Linear Programming LP: Linear Programming SP: Stochastic Programming SDP: Stochastic Dynamic Programming MINLP: Mixed Integer Non-Linear Programming GRASP: Greedy Randomized Adaptive Search Procedure NSGA-II: Non-dominated Sorting Genetic Algorithm II

From the table, we observe aspects of harvesting that were incorporated into the decisions in the reviewed papers. These include:

▪ Labor and/or routing/scheduling resources for harvesting: (Ferrer et al., 2008; Gómez-

Lagos et al., 2021; Mahdavimanshadi et al., 2024; Varas et al., 2020; Yao et al., 2023; Zuniga Vazquez et al., 2021).

▪ Perishability and quality which is managing short product lifespans and maintaining

freshness or quality (Fathollahi-Fard et al., 2023; Nagasawa et al., 2009; Varas et al.,

2020)

▪ Resource and technical constraints: these are limitations in machinery, storage,

processing, or field capacity (Herrera-Cáceres et al., 2017; Junqueira & Morabito, 2019; Khalilzadeh & Wang, 2022).

▪ Transportation and logistics coordination which is the integration of field operations with

processing plants or mills (Escallón-Barrios et al., 2022; Rocco & Morabito, 2016)

▪ Land management and zoning and/or spatial coordination between producers and

wholesalers (Albornoz et al., 2022)

▪ Uncertainty and stochasticity studies random factors such as weather, yield, or market

demand ((Albornoz & Vera, 2023; López et al., 2024; Mahdavimanshadi et al., 2024)

▪ Environmental and sustainability considerations related to minimizing waste, emissions,

and/or resource use (e.g., (Caicedo Solano et al., 2020; Chan et al., 2025) In terms of modeling approaches, most reviewed models use MILP that has the added advantage of a wide availability of solvers embedded in both academic and commercial software tools. LP is limited to models, in which variables were primarily quantities (Caixeta-Filho et al., 2000; Custodio et al., 2024; Hoogeveen et al., 2019; Rocco & Morabito, 2016; Widodo et al., 2006). Extensions such as SP models (Albornoz & Vera, 2023; López et al., 2024; Mahdavimanshadi et al., 2024) use the “now and then” and “recourse” approach in which different scenarios and their probabilities of occurrence are incorporated. These works were used primarily for weather and demand uncertainty. SDP is far less common with only two reviewed papers (Boussios et al., 2019; Zhang & Swaminathan, 2020) with applications that provide responses to weather events in crop planning (see table 1).

Regarding objective functions, most are related to economic metrics (cost minimization, profit maximization or net present value) but there are other functions related to quality (Fathollahi-Fard et al., 2023; Varas et al., 2020), waste (Chan et al., 2025), emissions (Caicedo Solano et al., 2020) and productivity (Y. Zhang & Swaminathan, 2020). Normally, these other objectives were combined with some economic objective in multi-objective optimization. As seen in the background section and in the literature review, although clearly roses share some characteristics with other flowers and agricultural products, their unique cultivation and harvesting characteristics make the decision making notoriously different:

▪ Roses in tropical countries are perennial crops whose harvesting cycles can be manually

adjusted by cutting and activating the stems. In the case of the rose industry, these cycles are adjusted to meet demand. Other perennial crops such as guayule have fixed harvesting seasons that cannot be adjusted.

▪ In the case of greenhouse cut roses in tropical countries (e.g. Ecuador and Colombia)

weather is not normally an issue. Temperatures, humidity, and luminosity do not vary significantly throughout the year and minor variations in these factors are controlled inside the greenhouses. Although we cannot say that weather does not affect flowering, variations primarily in daylight temperatures, are handled on site on a daily basis. Likewise, demand is known for the most part because the company works primarily under contracts with their customers.

▪ Unlike lilies and gladiolus, planting in roses takes place every ten or twelve years.

Predicting the varieties that will be sold in that time horizon is a decision at the strategic level. Likewise, the use of land and/or greenhouses is not an issue at this decision-making level.

▪ Although skilled labor can sometimes be limited, especially at peak times, the very high

perishability of the cut roses requires full labor availability at all stages (harvest, postharvest, shipping, etc.). We did not consider labor as a constraint in this research.

▪ In other agricultural products such as tomato, olive oil and palm oil, the processing is made

outside the farm, and some coordination is required between farms and processing plants; in rose production in Ecuador (and in Colombia), pest treatments, hydration and packing are made in the farm premises.

▪ Harvesting problems normally determine optimal timing and quantities of harvest for

crops, forests, or fisheries, maximizing total profit or yield while considering biological growth, perishability, capacity, and resource constraints. Unlike other agricultural products, the only harvest time in roses occurs when the flower is in bloom which is dependent on the initial cutting time. All quantities of flowers must be harvested when the flowers are in bloom. In other crops, harvesting is seasonal and quantities depend on the number of available plants.

The above literature review revealed that there exists a large body in the literature in a wide variety of agricultural planning and harvesting problems that contribute to our research. However, the literature also shows a significant gap in terms of cutting/harvesting models for rose and other perennial flowers. This research aims to close that gap. The following section illustrates our optimization model.

3.4

A Rose Production Model This section presents a MILP-based programming model aimed at defining a flower cutting schedule. The choice of MILP was due to several factors: (1) the availability of commercial software for future implementations makes it ideal for this project. (2) SP and DP can be prohibitive in terms of number of scenarios of combinations of demand and GDD curves. Also these models have been used primarily for strategic planning rather than for tactical planning as in this model. With the stochastic variations, the model may be easily intractable, and the benefit may not be clear as explained above.

Although initially tailored for the Freedom variety, the model is inherently flexible and can be adapted to other rose varieties based on their specific characteristics. This adaptability is achieved by adjusting parameters such as sales price, production costs, excess and shortage costs, production minimum quantities, GDD curves, and demand of which depend on the flower variety and the observed study period.

Now, we present the mathematical formulation of the model along with relevant considerations. The objective function of the farm is to maximize annual profits considering revenues, production costs, excess (dumps) and backorder costs. In this formulation, we use a discrete time approach. The decision variables of the model are mainly related to the (i) quantities of stems to cut at any period (week), (ii) the duration of the production cycles and (iii) the characteristics of perennial crops with continuous cycles in which every time that a stem is cut, a new cycle starts. To control the duration of the cycles, we use binary variables that establish the start and finish week of a given cycle. Since we also need to track the expected values of quantities of flowers that are harvested in a given week, we also use binary variables that specify whether there are stems that started their cycle at a given week. Parameters of the model include the GDD curves that determine the probabilities of bloom, price variations, customer demand, market seasonality, unit production costs, unit excess costs, and unit backorder costs.

Regarding constraints, these include logical, biological, capacity, and business types. Logical constraints define the relationship between the start and end of cycles and the activation of cycles. For example, a cycle that begins in one week must conclude in a future week and is considered active during all intermediate weeks. Biological constraints pertain to the GDD curve, the perishability of roses (which prevents inventory from being carried over between weeks), and their perennial nature. Capacity constraints establish that the number of stems cut cannot exceed the weekly availability on the farm, nor can it surpass the total annual production capacity. Finally, business constraints encompass the fulfillment of minimum quantity requirements and backorders.

3.4.1 Notation

Sets 𝑇= {1… 52} set of time periods (weeks) indexed in 𝑠, 𝑡, 𝑡′. Normally, 𝑠 is the start week, 𝑡 is the end week and 𝑡’ corresponds to a week in between.

Parameters 𝑐𝑝𝑡 = Unit production cost in week 𝑡 (USD 𝑠𝑡𝑒𝑚 ⁄

).

𝑐𝑜𝑡= Unit excess cost in week 𝑡 (USD 𝑠𝑡𝑒𝑚 ⁄

).

𝑐𝑢𝑡= Unit backorder cost in week 𝑡 (USD 𝑠𝑡𝑒𝑚 ⁄

).

𝑝𝑡= Unit sales price in week 𝑡 (USD 𝑠𝑡𝑒𝑚 ⁄

).

𝜓𝑡= Probability degree-days density curve in week 𝑡. 𝑚𝑡= Minimum quantities required production required in week 𝑡 (𝑠𝑡𝑒𝑚𝑠). 𝑑𝑡= Demand in week 𝑡 (𝑠𝑡𝑒𝑚𝑠).

𝑙𝑡 = % loss of flowers in week 𝑡. Includes those flowers sold in the local market and due to defects, mishandling, etc.

𝑎= available stems to cut which are calculated as the product of density of stems (𝑠𝑡𝑒𝑚𝑠 𝑚2 ⁄

)

times the planting area of the farm (𝑚2). 𝑛𝑐 = Average number of yearly production cycles (𝑐𝑦𝑐𝑙𝑒𝑠/𝑦𝑒𝑎𝑟).

Decision Variables 𝑋𝑠𝑡 = {1 if the production cycle starts at the beginning of week s and terminates at the beginning of week t

0 Otherwise

𝑌𝑠𝑡′𝑡 = {1 if the production cycle starts at the beginning of week s, terminates at the beginning of week t and is active

0 Otherwise

𝑈𝑠𝑡′ = {1 if the production cycle starts at the beginning of week s and is active in week 𝑡′

0 Otherwise

𝐻𝑡 = total stems harvested in week 𝑡. 𝑂𝑠= stems cut in week 𝑠.

𝑄𝑠𝑡′ = stems cut at the beginning of week 𝑠 that are active in week 𝑡′. “Active” stems are unavailable for cutting.

𝐸𝑡 = exported stems by the farm in week 𝑡. 𝐼𝑡 = Net inventory in week 𝑡 (𝑠𝑡𝑒𝑚𝑠). 𝐼𝑡 + = Excess inventory in week 𝑡 (𝑠𝑡𝑒𝑚𝑠). 𝐼𝑡 −= Backordered inventory in week 𝑡 (𝑠𝑡𝑒𝑚𝑠) 𝐴𝑡 = Available capacity in week 𝑡 (𝑠𝑡𝑒𝑚𝑠).

3.4.2 Model Formulation

This section presents our MILP model. Objective Function 𝑀𝑎𝑥 𝑃𝑟𝑜𝑓𝑖𝑡= ∑(𝑝𝑡𝑑𝑡 𝑡∈𝑇

- 𝑐𝑝𝑡𝐸𝑡 - 𝑐𝑢𝑡𝐼𝑡

− - 𝑐𝑜𝑡 𝐼𝑡 +)

Constraints ∑𝑋𝑠𝑡 𝑡>𝑠 ≤1 ∀𝑠 ∈ 𝑇

(1)

𝑋𝑠𝑡= 𝑌𝑠𝑡′𝑡 ∀ 𝑠, 𝑡′, 𝑡∈ 𝑇 such that 𝑠≤𝑡′, 𝑠 < 𝑡, 𝑡′ ≤𝑡

(2)

𝑈𝑠𝑡′ = 𝑌𝑠𝑡′𝑡 ∀𝑠, 𝑡′, 𝑡 ∈ 𝑇 such that 𝑠≤𝑡′, 𝑡′ < 𝑡 ∈ 𝑇

(3)

𝑄𝑠𝑡′ = Os 𝑈𝑠𝑡′ ∀ 𝑠, 𝑡′, 𝑡 ∈ 𝑇 such that 𝑠≤𝑡′, 𝑠 < 𝑡, 𝑡′ ≤𝑡

(4)

𝑂𝑠≤𝑎∑𝑋𝑠𝑡 𝑡

∀𝑠, 𝑡∈𝑇 such that 𝑠 < 𝑡

(5)

𝐻𝑡′ = ∑ψ𝑡′−𝑠+1𝑄𝑠𝑡′ 𝑠≤𝑡′

∀ 𝑡′ ∈ 𝑇

(6)

∑𝐻𝑡 ≤𝑎 𝑡∈𝑇 (1 −𝑙𝑡)𝑛𝑐 ∀𝑡∈ 𝑇

(7)

𝑚𝑡≤ 𝐻𝑡≤𝑂𝑡 ∀𝑡∈ 𝑇

(8)

𝐴𝑡= 𝑎+ 𝐻𝑡−∑𝑄𝑡𝑡′ 𝑡′

∀𝑡∈ 𝑇

(9)

𝑂𝑡≤𝐴𝑡 ∀𝑡∈ 𝑇

(10)

𝐼𝑡 = 𝐼𝑡 + - 𝐼𝑡 − ∀𝑡∈ 𝑇

(11)

𝐼𝑡 = 𝐻𝑡 - 𝑑𝑡 ∀𝑡∈ 𝑇

(12)

𝑑𝑡 = 𝐸𝑡 + 𝐼𝑡 − ∀𝑡 ∈𝑇

(13)

𝑋𝑠𝑡, 𝑌𝑠𝑡′𝑡 , 𝑈𝑠𝑡′ ∈ {0,1} ∀ 𝑠, 𝑡′, 𝑡∈ 𝑇 such that 𝑠≤𝑡′, 𝑠 < 𝑡, 𝑡′ ≤𝑡

(14)

𝐻𝑡, 𝑄𝑠𝑡, 𝐸𝑡, 𝐼𝑡 +, 𝐼𝑡 −≥ 0 ∀𝑠, 𝑡 such that 𝑠 < 𝑡∈𝑇

(15)

The objective function is the profit maximization that includes the constant revenue term ∑ 𝑝𝑡𝑑𝑡 𝑡 ∈ 𝑇 , the production and the excess and backorder costs. We could have dropped the constant

term and minimized the total cost but the farm’s management preferred profit maximization. The model has a total of 15 constraint sets, described as follows:

Constraint set 1 ensures that if a production cycle starts at period 𝑠, it must invariably terminate at some period 𝑡. Constraint 2 relates the binary variables 𝑋𝑠𝑡 and 𝑌𝑠𝑡′𝑡. This is, if a cycle starts in week 𝑠 and terminates in week 𝑡 then the cycle is active in week 𝑡′ given that 𝑠≤ 𝑡′ ≤𝑡. Similarly, constraint 3 relates variables 𝑌𝑠𝑡′𝑡 and 𝑈𝑠𝑡′. This is also logical consistency: if a cycle started in week 𝑠 and is active in week 𝑡′ then it must terminate at some week 𝑡. Constraint set 4 is the definition of 𝑄𝑠𝑡′ (stems active in week 𝑡′ whose production cycle started in week 𝑠) as the product of two variables 𝑂𝑠 (stems cut in week 𝑠) and 𝑈𝑠𝑡′ (if there is a cycle that started in week 𝑠 and is active in week 𝑡’). This is the quantity of active stems that are currently in production in week 𝑡′ whose cycle started in week 𝑠. Constraint set 5 relates variables 𝑋𝑠𝑡 and 𝑂𝑠. Essentially, every time that 𝑂𝑠 stems are cut, a new cycle starts that must terminate at some week 𝑡. This non-linear constraint can easily be linearized because it is well-known case of the product of one continuous and one binary variable. Constraint set 6 is the calculation of the stem production in week 𝑡′ and is contingent upon the probability density GDD curve. Constraint sets 7 and 8 impose limits on the harvesting quantities. Constraint set 7 establishes that the total number of harvested flowers in the year cannot exceed the total capacity of the farm. Constraint set 8 establishes that the harvested quantities in a week cannot either be below the predefined minimum (𝑚𝑡) or be above the cut quantities (𝑄𝑡) in the same week. Constraint set 9 defines the available quantities of stems that can be cut. This is the total number of available stems plus the harvested stems minus the total number of active stems in each week. Constraint set 10 imposes limits on the quantities of stems that can be cut. These quantities must be less or equal than the number of available stems. Constraint sets 11 to 13 represent the balance between inventory, sales, and the quantities harvested. Constraint set 11 defines a free variable as the difference between two non-negative variables; constraint set 12 is the equation of net inventory which is the harvested quantities less the demand. Notice that no inventory can be carried from the previous week because the product is perishable. Constraint set 13 is also an inventory balance constraint that equals the demand to the number of exported stems plus the expected backordered quantities. Constraint sets 14 and 15 correspond to the domain of the variables.

3.5

An illustrative real case As said above, the model used data from the María Farm located in the Cayambe municipality of Ecuador. The name of the farm was changed for confidentiality reasons. María Farm is one of the farms of a larger consortium with farms in Ecuador and Colombia and with a trading office in the United States. María Farm has an overall net production capacity of 29 million stems per year, consisting of Freedom and other colored roses. In the case of colored roses, the farm produces a range of 40 different varieties, which are grown in greenhouses. These flowers are primarily exported to the United States market. The flower sales are handled through various channels such as wholesalers and retailers such as supermarkets, and convenience stores. The Freedom variety was selected for the study, because it represents 42.3% of the cultivated area (11.5 hectares out of 27.2 hectares) in the María Farm. The production capacity of the freedom rose is roughly 12.5 million stems per year. Additionally, the model incorporates minimum weekly production levels. These minimums are a commitment of the sales department to its customers. Meeting this target

is critical for the business, as failure to do so forces the company to purchase flowers from third parties, which directly erodes profit margins. These minimums represent roughly 56% of the gross annual production.

María Farm's financial records only account for revenue from flowers produced on-site. While back-orders are managed by the US trading office and are not included in the farm's own revenue or cost statements, we incorporated them into the objective function to capture the business's total profit. Operationally, when the farm cannot meet demand, the US office sources the required flowers at market prices.

3.6 Computational experience and analysis

3.6.1

Data collection Data was collected from the records of the farm in 2021. These were the only data authorized by the farm for use by our research team. The production cost corresponds to the labor cost (at greenhouses and in postharvest) plus supplies (fertilizers, pesticides, hydration and packaging materials). The backorder cost corresponds to the market price of flowers at the trading offices in the US; the excess cost is the unit production cost minus the sum of the labor cost in postharvest, the cost of packing materials and the unit hydration cost. The GDD curve of the Freedom rose variety was also provided by the farm, discretized and transformed into a time-based curve. Data of sales (demand) were retrieved from the financial records of the farm. In Ecuador, flower export prices are public and can be obtained from the United States Department of Agriculture (USDA) website (www.usda.gov). Technical data such as the GDD curves, cultivated area and productivity were collected from the Statistics and Forecasting Department. The minimum committed quantities were provided by the Commercial Department.

The actual results of the business for the 2021 year are shown in Table 2, where at the end of the year, María Farm reported a gross margin of USD 1,032,551. The farm sold around 12.15 million stems of the Freedom variety rose.

Table 5. Results María Farm-2021 (Business as usual) P&L lines Current value

1. Annual sales

12,152,272 stems

2. Farm Sales (excludes backorders)

USD 3,337,987

3. Supplies cost (fertilizers, pesticides, etc.)

USD 872,238

4. Labor cost

USD 1,411,877

5. Operational cost (productive)

USD 2,284,115 (3+4)

6. Operational costs (vegetative)

USD 21,321

7. Total Operational cost

USD 2,305,436 (5+6)

8. Farm Gross Margin

USD 1,032,551 (2-7)

*Data taken from the accounting records of María Farm.

It is important to note that in the financial results of María Farm, the gross margin calculation does not account for excess (dumps) and backorders. If the farm were to account for these extra costs in the financial results, the gross margin would have been lower than the reported figures.

3.6.2

Model results The model was implemented in the IBM ILOG CPLEX 22.1.1 IDE. The problem instance for the case study consists of 204,284 constraints and 194,417 variables. The computations were performed on a system running the Windows 11 operating system, equipped with an 11th Gen Intel® Core™ i7-1135G7 processor (@ 2.40GHz), 16 GB of RAM, and a quad-core CPU capable of executing eight threads simultaneously. IBM ILOG CPLEX was set to use its default parallel mode, utilizing all available threads. The solution process was configured to terminate upon reaching an optimality gap of 2% or less, a condition most runs satisfied in approximately 45 minutes. A comparison between the model's results and the actual figures from María Farm is presented in Table 6.

Table 6. Results of the María Farm-2021 (Real versus model results) Item Real value Model Results Model versus Real (%)

1. Total exports (stems)

12,152,272

12,910,746

6.2%

2. Excess (dumps) (stems)

698,000

507,005

-27,3%

3. Backorders (stems)

1,227,000

590,327

-51,8%

4. Farm Sales (USD)

3,337,987

3,694,900

10.7%

5. Production Cost (USD)

2,305,436

2,413,676

5,7%

6. Excess cost (USD)

109,655

81,121

-26,0%

7. Backorder Cost (USD)

513,213

130,974

-74,4%

8. Farm

Gross Margin

(USD)

1,032,551

1,259,993

+22,0%

The cutting schedule curve presented in

Figure 4 is similar to the actual schedule reported by the farm in 2021 although the peak times were different.

The model was extensively validated with the farm's senior management. For María Farm, its implementation led to a significant profit increase, generating an annual profit of approximately USD 1,259,993—a 22% increase over the actual profit of USD 1,032,550, on a production volume of 12.9 million stems. This improvement is primarily attributed to a 51.8% reduction (approximately 636,673 stems) in back-orders, coupled with a 27.3% reduction in dumped stems. Although the model's outputs are realistic, the results of executing this plan may vary due to fluctuations in harvested quantities caused by weather and market conditions. The model's performance can be explained by three key factors:

• It meets the minimum production quantities to meet demand, reducing backorders. This is

not always achieved with the actual production plan.

• It recommends increasing production for the Mother's Day period (weeks 15-19), thereby

boosting revenue, as shown in Figure 3.

• It improves the balance between excess and shortages year-round.

Figure 4 Cutting curve (harvested stems) for María Farm. Figures in thousands of stems. In essence, the model achieved a 34% reduction in the backorder-to-excess ratio. This was accomplished mainly by minimizing backorders, which have a far greater financial impact than inventory dumps.

Figure 5 presents a Gantt chart of the 52-week production cycles at María Farm, where each cycle lasts between 10 and 14 weeks.

Figure 5 Production Cycles Timeline in María’s farm The chart shows the overlapping cycles; for instance, Cycle 0 starts in week 1 and finishes in week 11, while Cycle 1 starts in week 2 and finishes in week 14. This overlap shows the weekly harvesting of distinct stems, meaning the batch cut in one week is distinct from the batches cut in previous or subsequent weeks.

3.6.3 Sensitivity Analysis

To test the model's sensitivity to key inputs, we conducted an analysis focusing on parameters critical to business margins. Minimum sales quantities, unit production costs, and selling prices were recommended to carry out this analysis. The model's financial outcomes were highly sensitive to variations in these parameters within predefined ranges recommended by senior managers. All scenarios were compared against a baseline, which was defined by the initial farm parameters as explained in Section 6.1.

The first analysis evaluates the impact of adjusting minimum quantity commitments. One strategy is to increase these commitments to ensure demand is met. However, this can result in higher production volumes during low-price seasons, thereby reducing potential revenue. The opposite strategy, decreasing minimums, allows the model to produce larger quantities at highprice peak seasons to maximize revenue. The primary risk of this approach is a significant increase in backorders during the low season. This, in turn, could force the farm to make unfavorable spot purchases on the open market to fulfill its contractual obligations. Table 7. Sensitivity Analysis: Influence of Minimums’ shows the results. Each column shows the figures for a percentage increase in the prescribed minimums. For example, the column +2% minimums denotes an

increase of 2% in the minimum quantities throughout the year. The baseline scenario corresponds to the results presented in Table 6. ¡Error! No se encuentra el origen de la referencia. and 6 also illustrate the results of the sensitivity analysis. Table 7. Sensitivity Analysis: Influence of Minimums’ Variations

Actual

-5%

-2%

Baseline +2% +5% Gross profit

(USD)

1,032,551

1,381,223

1,307,699

1,259,993

1,211,367

1,138,351

Backorder costs (USD)

513,213

129,503

121,881

130,974

127,123

130,385

Excess costs

(USD)

109,655

78,703

88,995

81,121

77,084

78,314

Exported stems (stems)

12,152,27

2

12,931,12

7

12,949,96

6

12,910,74

6

12,938,07

3

12,911,61

3

Excess (stems)

698,000

491,892

556,219

507,005

481,772

489,462

Shortages (stems)

1,227,000

569,947

551,107

590,327

569,756

589,463

Figure 6. Financial results vs minimums’ changes

0

200

400

600

800

1000

1200

1400

Actual

-5%

-2%

Baseline +2% +5% (USD in thousands) Financial results vs Changes in Minimum Quantities Gross profit (USD) Backorder costs (USD) Excess costs (USD)

Figure 7. Overstock (dumps) and shortages vs minimums’ changes.

In the first sensitivity analysis, where we varied the minimum quantity values within a range of -5% to +5%, we found that the highest profit margin was achieved by reducing the minimum quantities by 5%. This change increased the gross margin by 33% compared to the actual business results. The model suggests this increase is primarily due to reallocating stems to peak seasons, enabling them to be sold at higher prices and increasing overall farm revenue by 13.3%. This 5% reduction represents the lower limit of what the farm is willing to implement. Further reductions risk the loss of contracts with major buyers, which could significantly impact sales during the low season. The model also consistently outperforms current practice in managing both excess inventory and shortages. Excess costs under the model were between USD 77,158 and USD 87,035, compared to USD 109,655 of the current practice. Similarly, backorder costs were dramatically lower, ranging from USD 125,657 to 136,341 versus the current cost of USD 513,213. Next, we conducted a sensitivity analysis regarding the production cost, as outlined below. Table 5, and ¡Error! No se encuentra el origen de la referencia. and 8 illustrate the above results. Table 5. Sensitivity Analysis: Influence of Production Cost

Actual

-5%

-2%

Baseline +2% +5% Gross profit

(USD)

1,032,551

1,380,053

1,307,868

1,259,993

1,210,180

1,138,657

Backorder costs (USD)

513,213

125,657

136,341

130,974

131,980

127,259

Excess costs

(USD)

109,655

77,158

82,903

81,121

80,162

87,035

Total exports (stems)

12,152,27

2

12,936,50

8

12,897,30

3

12,910,74

6

12,914,20

5

12,927,46

9

Excess (stems)

698,000

507,620

528,716

507,005

491,186

518,067

Shortages (stems)

1,227,000

564,565

603,770

590,327

586,868

573,064

0

200

400

600

800

1000

1200

1400

Actual

-5%

-2%

Baseline +2% +5% Stems (in thousands) Excess and Shortages vs Changes in the Minimum Quantities Excess (stems) Shortages (stems)

In this second sensitivity analysis, the unit production cost parameter varied from -5% to +5% with respect to the actual production cost, leaving everything else unchanged. The sensitivity analysis shows, as expected, that the highest gross margin is achieved when production costs are reduced by -5%. We observe that these results, in terms of gross margin, are not significantly different from those obtained in the sensitivity analysis conducted with a -5% reduction in minimums.

Figure 8. Gross profit vs production cost variations.

Figure 9. Excess (dumps) and shortages vs production cost variations. We can infer that reducing minimum quantities by -5% is almost equivalent to reducing production costs by the same proportion (-5%). However, in practice, reducing production costs by these percentages is very difficult, as labor is the largest component of production costs, and this cost increases every year in proportion to the inflation rate. As in the minimum quantities case, the reduction in excess (dumps) and shortages was also significant.

0

200

400

600

800

1000

1200

1400

Actual

-5%

-2%

Baseline +2% +5% (USD in thousands) Financial results vs Changes in Unit Production Cost Gross profit (USD) Backorder costs (USD) Excess costs (USD)

0

200

400

600

800

1000

1200

1400

Actual

-5%

-2%

Baseline +2% +5% Stems (in thousands) Excess and Backorders vs Changes in Unit Production Cost Excess (stems) Backorders (stems)

In our third and last sensitivity analysis, we varied the unit price parameter as outlined below. Table

6

shows the results.

Figure 10 and 10 also illustrate the analysis. In this sensitivity analysis, the unit price parameter varied from -10% to +10%, leaving everything else unchanged as in previous analysis. Table 6. Sensitivity Analysis: Influence of Selling price

Actual

-10%

-5%

Baseline

5%

10%

Gross profit

(USD)

1,032,551

897,945

1,075,281

1,259,993

1,445,481

1,625,363

Backorder costs (USD)

513,213

137,303

121,754

130,974

146,166

133,838

Excess costs

(USD)

109,655

76,675

77,068

81,121

90,165

81,389

Total exports (stems)

12,152,27

2

12,811,35

9

12,922,28

8

12,910,74

6

12,889,29

4

12,962,57

8

Overstock (stems)

698,000

478,598

481,678

507,005

563,530

508,680

Understock (stems)

1,227,000

689,714

578,785

590,327

611,779

538,495

0

200

400

600

800

1000

1200

1400

1600

Actual

-10%

-5%

Baseline

5%

10%

(USD in thousands) Financial results vs Changes in Unit Price Gross profit (USD) Backorder costs (USD) Excess costs (USD)

Figure 10. Gross profit vs production price changes.

Figure 11. Overstock (dumps) and shortages vs price changes A critical consideration is that a uniform +10% price increase affects not only seasonal and standard prices but also elevates backorder costs, as these are tied to market prices. It should be noted that while such price adjustments are feasible in this industry, they typically occur over the long term; for context, market prices have risen approximately 8% over the past four years. It is important to notice that in the rose production industry, demand during peak seasons such as Valentine’s Day and Mother’s Day is highly inelastic, meaning that it is not significantly affected by changes in the selling price. Essentially, in these two peak seasons, all that is produced is sold. In the low season, prices are mostly established by customers.

The model identifies a reduction in minimum commitments as the most effective lever for improving margin. A -5% reduction in minimums yields a gross margin of USD 1,381,223. To achieve an equivalent margin through price increases alone, the business would need to implement a uniform price adjustment of +3.27%. Similarly, this result is comparable to a -5% reduction in

0

200

400

600

800

1000

1200

1400

1600

Actual

-10%

-5%

Baseline

5%

10%

(USD in thousands) Financial results vs Changes in Unit Price Gross profit (USD) Backorder costs (USD) Excess costs (USD)

0

200

400

600

800

1000

1200

1400

Actual

-10%

-5%

Baseline

5%

10%

Stems (in thousands) Excess and Backorders vs Changes in Unit Price Excess (stems) Backorders (stems)

production costs. Therefore, the model demonstrates that a slight reduction in minimum commitments is strategically equivalent to a modest price increase or a significant gain in operational efficiency.

Regarding system behavior, the volume of excess stems gradually increased as prices varied from -10% to +5%, but saw a sharp decrease at a +10% price increase. Conversely, backorders peaked when unit prices fell by -5% and -10% but reached their minimum at a +10% price increase. This occurs because the associated increase in backorder penalty costs at higher price levels makes fulfilling demand through production more economically advantageous than incurring those penalties.

3.6.4 Managerial insights

From a managerial standpoint, the results highlight how optimization-based planning can transform rose production from an empirical process into a data-driven decision system. Our model provides guidance on how to define cutting schedules, helping managers visualize the trade-offs between profitability, reliability (meeting demand), and risks (backorders). For instance, the sensitivity analyses demonstrate that adjustments in weekly minimum quantities (–5%) or production costs (–5%) can yield profit increases comparable to increments in price (+3%). The model also quantifies the economic implications of dumps and backorders, giving managers objective measures to evaluate how to adjust production cycles and negotiate sales commitments (minimums). The incorporation of GDD curves into the model allows managers to combine biological aspects of flower production with commercial aspects. In a broader view, the approach offers a framework for long-term capacity planning and resource use optimization in floriculture, that can eventually help the farm negotiate quantities and prices with customers. Clearly, the model is highly dependent on accurate data of demand, prices and costs. Likewise, its implementation requires a mature business model in which production plans are executed with. In this industry, many farms plan their production on a weekly basis and keep reacting to changes in demand (same-day orders).

3.7

Conclusions This research presents a novel contribution by providing a mathematical model that enables rose farms to plan their production with mathematical models. The proposed model considers key business parameters, such as the farm's production capacity, phenological stages, degree day curves, demand, prices, production costs, as well as costs associated with excess inventory and backorders. This can be achieved by (i) anticipating the number of stems that need to be produced annually to meet demand, (ii) specifying the dates when the product should be available for delivery, (iii) specifying how production should be distributed throughout the year to address seasonal and every-day needs, and (iv) determining the producer's margin at the end of each year. Additionally, it enables producers to define, well in advance, the number of stems available to promise during negotiations with their customers. The model can be adapted to other rose varieties and to other species in which the flower cycle times can be varied This research responds to the questions that flower growers often ask: (i) what is the best cutting plan that achieves the highest profits, with timely deliveries and in the right quantities, throughout

the 52 weeks of the year? (ii) What should be the minimum quantities to harvest weekly to guarantee fulfilling contracts in the low season? (iii) What are the quantities of unavoidable dumps and backorders resulting from this optimized cutting plan. An extra identified benefit from applying the new tool was the visualization of production planning. The results generated by the model enable the company to develop a visible business plan based on real and farm-specific data, facilitating better decision-making through the early identification of problems and possible solutions, such as purchasing flowers in advance, running promotions to reduce dumps, evaluating prices, and so forth.

It is evident that the model adopts an optimistic perspective, as it assumes both deterministic demand and flowering cycles. Rose flowering is complex and it depends on external factors. As said, the actual harvested quantities may deviate from those predicted and the excess and backorder stems may be higher than those computed by the model. However, since the María Farm has modern control technologies, we can affirm that our model results are accurate enough for decision making. Incorporating stochastic features can complicate unnecessarily the model as discussed above.

There are several barriers for the implementation of the model. First, a commercial software license is required, this can be costly and most farms would not be willing to pay for a license that would be used a few times per year. Given the size of the model, open source or free software is not an option due to size limitations. Second, expert personnel in both the flower business and in mathematical models are needed to run and maintain these models and this is not a combination easy to find. Developing a realistic MILP model required extended visits to rose farms both in Colombia and Ecuador for a full understanding of the business and of the agronomical aspects of flowers. The combined efforts of experts in math optimization, rose growth cycles and in business models made possible the development of this paper. The authors of this paper plan to introduce the tool through industry associations such as Expoflores in Ecuador and Asocolflores in Colombia.

Future developments will focus on other aspects such as labor constraints and on sustainable practices. In future research, the model can also be combined with Industry 4.0 technologies for forecasting (Mantilla et al., 2025b). For example, image processing with advanced technologies such as Deep Learning can be used to better predict the phenological stages; Internet of Things (IoT) can be used in combination with wireless sensors to control humidity and temperature which have direct impact also in the flower bloom. Although IoT with wireless sensors for monitoring are currently in use in the greenhouses of María Farm, there is still a gap to really adopt these technologies and models.

Acknowledgements The authors wish to express their gratitude to the personnel of María Farm for their valuable data and insights for the model.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the author(s) used ChatGPT and DeepSeek to solely to improve grammar and style. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Fernando Mantilla reports article publishing charges was provided by Falcon Farms SAS. Fernando Mantilla reports a relationship with Falcon Farms SAS that includes: employment. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Chapter 4

4. Conclusions and Future Research

This research has presented (i) an overview and a discussion of the technologies of Industry 4.0 in floriculture and the challenges and barriers they face for their implementation and, (ii) a datadriven optimization framework that links with these technologies. This thesis shows potential tangible benefits to the floriculture sector, as it not only identifies the application of the I4.0 technologies across the various links of the flower production chain, but also presents a conceptual framework that outlines a roadmap for implementation. These technologies are primarily focused on core processes such as planting, harvesting, and product distribution in the context of flower cultivation in emerging countries such as Colombia and Ecuador. Parts of this research were presented in one of the world’s most important floriculture events, SIFLOR, which took place in Ecuador in November 2025.

The foundations of a data-driven optimization framework were illustrated through a mathematical model applied to a case study in Ecuador. Beyond the modeling approach, this research presents practical strategies, and potential issues that may arise, required before its full implementation. Although the model is only a prototype, some changes in the production and planning processes have already been made in the farm, based on its results. The model allows flower growers to gain a clear and early view of the projected economic results (specifically, profit margin) and provides insight into resource requirements based on the planned distribution of harvests, offering a better understanding of expected excess levels and potential needs for flower purchases and other inputs required to meet customer demand.

We recommend that the implementation process of the model described above incorporates

structured data validation protocols, involving close coordination between the farm’s technical staff and commercial team. To reduce the risk of data inconsistencies and manual errors, it is crucial to implement automated systems for data collection and real-time updates, either through the integration of IoT-based sensors or through direct synchronization with the producer’s ERP system. Additionally, to address challenges related to real-time data availability and user input reliability, we need the development of an intuitive user interface that allows stakeholders to validate, modify, and approve input variables before running the model. This would enhance both the model’s accuracy and its usability in operational decision making. Future Research Labor Challenges in Emerging Markets One of the most critical challenges that the floriculture sector in emerging countries faces is the availability of labor for flower production. In recent years, the sector has experienced a growing shortage of field labor—an issue that threatens the long-term sustainability of agricultural operations. In countries like Colombia and Ecuador, it is becoming increasingly difficult to attract and retain labor willing to perform tasks related flower cultivation, post-harvest and logistics. Labor availability is a pressing issue that must be addressed within the next decade. Therefore, new research must focus on how the I4.0 technologies can help eliminate activities that impact labor costs in flower cultivation, such as inspection tasks, stem counting, temperature measurements at different points in the supply chain, and field monitoring, not only from the technical but also from the organizational perspective.

Modeling Rose production planning remains an underexplored topic; therefore, the analysis and development of new models applied to production planning and the creation of cutting schedules could be further deepened. Future work could incorporate the uncertainty associated with climate forecasts and the impact of plant health on flower crop productivity. Additionally, flower cutting times and other labor activities both of which are stochastic in nature should be considered, as well as the inclusion of social welfare indicators and other aspects of sustainability. All these aspects could enrich and complement the results obtained in this study.

The MILP model can also be extended by incorporating labor efficiency parameters that allow simulating the number of hours required to produce a given harvest volume and converting these values into the optimal number of workers needed to perform production operations in a standardized manner. This approach would help establish crop efficiency indicators, directly contributing to productivity, operational results, and business profit. Technologies Other lines of future research will be the extensions of the use of these technologies to other stages of the value chain, including business planning, pricing strategies, management and administration of crops and additional initiatives aimed at maximizing performance and ensuring long-term business sustainability. This can be achieved through combinations of hardware-based technologies combined with decision-support tools such as the model proposed in this research. For example, Blockchain is unexplored technology in the flower sector: this technology can

provide significant value to the export flower sector by enhancing traceability, transparency, coordination, and trust across a highly fragmented and perishable supply chain. The flower industry, particularly in exporting countries such as Colombia and Ecuador, involves multiple stakeholders: farms, transporters, logistics operators, airlines, customs authorities, importers, wholesalers, and retailers. The perishability of flowers and the importance of maintaining the cold chain make information integrity and real-time coordination critical. Blockchain can contribute in several ways: end-to-end traceability, cold chain traceability and smart contracts. However, its prohibitive costs and the need for full integration are barriers for its implementation. Research is critical to this. The same can be said for Big Data, Cloud Computing, Robotics and so forth. The range of potential solutions is broad, and we believe that future research in this field will open new business model opportunities, helping to address many of the challenges identified in this thesis.

References Chapter 2

1. de Carvalho, P.R.V., Naoum-Sawaya, J., Elhedhli, S.: Blockchain-Enabled supply chains: An

application in fresh-cut flowers.

Appl Math Model.

110,

841–858

(2022).

https://doi.org/10.1016/j.apm.2022.06.011

2. Adamides, G.: A review of climate-smart agriculture applications in Cyprus. Atmosphere

(Basel). 11, (2020). https://doi.org/10.3390/ATMOS11090898

3. Hassan, S.I., Alam, M.M., Illahi, U., Al Ghamdi, M.A., Almotiri, S.H., Su’ud, M.M.: A

Systematic Review on Monitoring and Advanced Control Strategies in Smart Agriculture. IEEE Access. 9, 32517–32548 (2021). https://doi.org/10.1109/ACCESS.2021.3057865

4. Lezoche, M., Panetto, H., Kacprzyk, J., Hernandez, J.E., Alemany Díaz, M.M.E.: Agri-food 4.0:

A survey of the Supply Chains and Technologies for the Future Agriculture. Comput Ind. 117,

103187 (2020). https://doi.org/10.1016/j.compind.2020.103187

5. Van Hentenryck, P., Dalmeijer, K.: AI4OPT: AI Institute for Advances in Optimization. AI Mag.

45, 42–47 (2024). https://doi.org/10.1002/aaai.12146

6. Pekkerieta, E.J., Van Henten, E.J., Campen, J.B.: Contribution of innovative technologies to

new developments in horticulture. (2015)

7. Et-taibi, B., Abid, M.R., Boufounas, E.-M., Morchid, A., Bourhnane, S., Abu Hamed, T.,

Benhaddou, D.: Enhancing water management in smart agriculture: A cloud and IoT-Based smart irrigation system.

Results in Engineering.

22,

102283

(2024).

https://doi.org/10.1016/j.rineng.2024.102283

8. Kamyab, H., Khademi, T., Chelliapan, S., SaberiKamarposhti, M., Rezania, S., Yusuf, M.,

Farajnezhad, M., Abbas, M., Hun Jeon, B., Ahn, Y.: The latest innovative avenues for the utilization of artificial Intelligence and big data analytics in water resource management. Results in Engineering. 20, 101566 (2023). https://doi.org/10.1016/j.rineng.2023.101566

9. Heemskerk, C., Boode, A.H., Arntzen, P., Fesselet, L.: HiPerGreen: Greenhouse crop scouting

by a UAS: Lessons learned from cross-domain cooperation in applied research. In: 2020 21st International Conference on Research and Education in Mechatronics, REM 2020 (2020)

10. Wani, M.A., Din, A., Nazki, I.T., Rehman, T.U., Al-Khayri, J.M., Jain, S.M., Lone, R.A., Bhat,

Z.A., Mushtaq, M.: Navigating the future: exploring technological advancements and emerging trends in the sustainable ornamental industry. Front Environ Sci. 11, 1188643 (2023). https://doi.org/10.3389/fenvs.2023.1188643

11. van Henten, E.J.: Greenhouse mechanization: state of the art and future perspective. Acta

Hortic. 710, 55–70 (2006). https://doi.org/10.17660/ActaHortic.2006.710.3

12. Hatfield, J.L., Prueger, J.H.: Temperature extremes: Effect on plant growth and development.

Weather Clim Extrem. 10, 4–10 (2015). https://doi.org/10.1016/j.wace.2015.08.001

13. Juroszek, P., Von Tiedemann, A.: Plant pathogens, insect pests and weeds in a changing global

climate: A review of approaches, challenges, research gaps, key studies and concepts. Journal of Agricultural Science. 151, 163–188 (2013). https://doi.org/10.1017/S0021859612000500

14. Fiorucci, A.S., Fankhauser, C.: Plant Strategies for Enhancing Access to Sunlight, (2017)

15.

Monroy, N., Cure, J.R., Pérez, I., Cure, J.R.: Estudio de la variabilidad en el clima y la producción de rosas en la sabana de Bogotá. Revista de Ingeniería. 14, 38–43 (2014)

16. Grand View Research: Cut Flowers Market Analysis 2018-2030. (2024)

17. Market Analysis Report: Floriculture Market-Global Industry Analysis and Forecast. (2023)

18. Dole, J.M.., Wilkins, H.F..: Floriculture: principles and species. 1023 (2005)

19. Reid, M.S., Jiang, C.-Z.: Postharvest Biology and Technology of Cut Flowers and Potted

Plants. Hortic Rev (Am Soc Hortic Sci). 40, (2012)

20. Jawaharlal, M., Ganga, M., Padmadevi, K., Jegadeeswari, V., Karthikeyan, S.: A technical

guide on carnation. (2010)

21. Higaki, T., Rasmussen, H.P., Carpenter, W.J.: A study of Some Morphological and Anatomical

Aspects of Anthurium andreanum Lind., http://hdl.handle.net/10125/25543, (1984)

22. Pérez, I., Cure, J.R., Monroy, N.: Modelo de predicción y manejo de cultivos de rosas. Revista

de Ingeniería. 15, 18 (2002)

23. Mejía, G., Sarmiento, A., Guerrero, W.: Truck and Cross-Docking Scheduling in the Flower

Export Sector in Colombia. Transp Res Rec. 03611981221130332 (2022)

24. Mejía, G., Guerrero, W., Sarmiento, A., Serrano, N., Sarmiento, M., Sánchez, C.: Delivery of

Perishable Export Products in Smart Cities: A Case Study in Bogotá (Colombia). Procedia Manuf. 39, 1946–1952 (2019). https://doi.org/10.1016/j.promfg.2020.01.237

25. Sepehriar, A., Eslamipoor, R.: An economical single-vendor single-buyer framework for

carbon emission policies.

Journal of Business Economics.

94,

927–945

(2024).

https://doi.org/10.1007/s11573-024-01192-w

26. Tsai, W.-C., Shen, C.-W.: Using a smart contract for the floral supply chain. Asia Pacific

Management Review. (2024). https://doi.org/10.1016/j.apmrv.2023.12.004

27. Ahumada, O., Villalobos, J.R.: Operational model for planning the harvest and distribution of

perishable agricultural products.

Int J Prod Econ.

133,

(2011).

https://doi.org/10.1016/j.ijpe.2011.05.015

28. Sekaran, K., Meqdad, M.N., Kumar, P., Rajan, S., Kadry, S.: Smart agriculture management

system using internet of things. Telkomnika (Telecommunication Computing Electronics and Control). 18, 1275–1284 (2020). https://doi.org/10.12928/TELKOMNIKA.v18i3.14029

29. Yang, J., Sharma, A., Kumar, R.: IoT-based framework for smart agriculture. International

Journal of Agricultural and Environmental Information Systems. 12, 1–14 (2021). https://doi.org/10.4018/IJAEIS.20210401.oa1

30. Hosny, K.M., El-Hady, W.M., Samy, F.M.: Technologies, Protocols, and applications of

Internet of Things in greenhouse Farming: A survey of recent advances, (2024)

31. Mana, A.A., Allouhi, A., Hamrani, A., Rahman, S., el Jamaoui, I., Jayachandran, K.:

Sustainable AI-based production agriculture: Exploring AI applications and implications in agricultural practices.

Smart Agricultural Technology.

7,

(2024).

https://doi.org/10.1016/j.atech.2024.100416

32. Morchid, A., Oughannou, Z., Alami, R. El, Qjidaa, H., Jamil, M.O., Khalid, H.M.: Integrated

internet of things (IoT) solutions for early fire detection in smart agriculture. Results in Engineering. 24, 103392 (2024). https://doi.org/10.1016/j.rineng.2024.103392

33. Morchid, A., Jebabra, R., Khalid, H.M., El Alami, R., Qjidaa, H., Ouazzani Jamil, M.: IoT-

based smart irrigation management system to enhance agricultural water security using embedded systems, telemetry data, and cloud computing. Results in Engineering. 23, 102829 (2024). https://doi.org/10.1016/j.rineng.2024.102829

34. Rosero-Montalvo, P.D., Erazo-Chamorro, V.C., López-Batista, V.F., Moreno-García, M.N.,

Peluffo-Ordóñez, D.H.: Environment monitoring of rose crops greenhouse based on autonomous vehicles with a wsn and data analysis. Sensors (Switzerland). 20, 1–15 (2020). https://doi.org/10.3390/s20205905

35. Rodríguez, S., Gualotuña, T., Grilo, C.: A System for the Monitoring and Predicting of Data in

Precision Agriculture in a Rose Greenhouse Based on Wireless Sensor Networks. In: Procedia Computer Science. pp. 306–313. Elsevier B.V. (2017)

36. Park, A., Li, H.: The effect of Blockchain technology on supply chain sustainability

performances. Sustainability (Switzerland). 13, 1–18 (2021). https://doi.org/10.3390/SU13041726

37. What Is Artificial Intelligence (AI)? | IBM, https://www.ibm.com/think/topics/artificial-

intelligence

38. Sun, X., Li, Z., Zhu, T., Ni, C.: Four-dimension deep learning method for flower quality grading

with depth information.

Electronics (Switzerland).

10,

(2021).

https://doi.org/10.3390/electronics10192353

39. Mann, H.M.R., Iosifidis, A., Jepsen, J.U., Welker, J.M., Loonen, M.J.J.E., Høye, T.T.:

Automatic flower detection and phenology monitoring using time-lapse cameras and deep learning. Remote Sens Ecol Conserv. 8, 765–777 (2022). https://doi.org/10.1002/rse2.275

40. Estrada, J.S., Vasconez, J.P., Fu, L., Cheein, F.A.: Deep Learning based flower detection and

counting in highly populated images: A peach grove case study. J Agric Food Res. 15, (2024). https://doi.org/10.1016/j.jafr.2023.100930

41. Wang, X. (Annie), Tang, J., Whitty, M.: DeepPhenology: Estimation of apple flower phenology

distributions based on deep learning.

Comput Electron Agric.

185,

(2021).

https://doi.org/10.1016/j.compag.2021.106123

42. Ahmed, S., Nicholson, C.E., Rutter, S.R., Marshall, J.R., Perry, J.J., Dean, J.R.: Use of an

unmanned aerial vehicle for monitoring and prediction of oilseed rape crop performance. PLoS One. 18, e0294184 (2023). https://doi.org/10.1371/journal.pone.0294184

43. Zhang, T., Vail, S., Duddu, H.S.N., Parkin, I.A.P., Guo, X., Johnson, E.N., Shirtliffe, S.J.:

Phenotyping Flowering in Canola (Brassica napus L.) and Estimating Seed Yield Using an Unmanned Aerial Vehicle-Based Imagery.

Front Plant Sci.

12,

686332

(2021).

https://doi.org/10.3389/fpls.2021.686332

44. Bumpus, W.: NIST Cloud Computing Standards Roadmap. NIST Cloud Computing Standards.

(2013)

45. Zhao, R., Wang, L., Zhang, X., Zhang, Y., Wang, L., Peng, H.: A OneM2M-Compliant Stacked

Middleware Promoting IoT Research and Development. IEEE Access. 6, 63546–63559 (2018). https://doi.org/10.1109/ACCESS.2018.2876197

46. Wang, S., Liu, H., Qin, X., Dai, J., Liu, J.: Cherry blossom and ginkgo leaf coloration

phenology dataset of China from 2009 to 2019 extracted from big data. Geosci Data J. (2023). https://doi.org/10.1002/gdj3.231

47. Sauer, P.C., Seuring, S.: How to conduct systematic literature reviews in management research:

a guide in 6 steps and 14 decisions. Review of Managerial Science. 17, 1899–1933 (2023). https://doi.org/10.1007/s11846-023-00668-3

48. Kamaruddin, F., Malik, N.N.N.A., Murad, N.A., Latiff, N.M.A., Yusof, S.K.S., Hamzah, S.A.:

IoT-based intelligent irrigation management and monitoring system using arduino. Telkomnika (Telecommunication Computing Electronics and Control).

17,

2378–2388

(2019).

https://doi.org/10.12928/TELKOMNIKA.v17i5.12818

49. Bunpalwong, M., Rukhiran, M., Netinant, P.: Improving marigold agriculture with an IoT-

driven greenhouse irrigation management control system. Bulletin of Electrical Engineering and Informatics. 12, 3817–3825 (2023). https://doi.org/10.11591/eei.v12i6.6300

50. Burhan, I.M., Ali, Q.A., Hussein, I.S., Jaleel, R.A.: Mobile-computer Vision Model with Deep

Learning for Testing Classification and Status of Flowers Images by using IoTs Devices. J Wirel Mob Netw Ubiquitous Comput Dependable Appl.

14,

82–94

(2023).

https://doi.org/10.58346/JOWUA.2023.I1.007

51. Patel, D.H., Shah, K.P., Gupta, R., Jadav, N.K., Tanwar, S., Neagu, B.C., Attila, S., Alqahtani,

F., Tolba, A.: Blockchain-Based Crop Recommendation System for Precision Farming in IoT Environment. Agronomy. 13, (2023). https://doi.org/10.3390/agronomy13102642

52. Atmaja, A.P., Hakim, A.E., Wibowo, A.P.A., Pratama, L.A.: Communication systems of smart

agriculture based on wireless sensor networks in IoT. Journal of Robotics and Control (JRC). 2, 297–301 (2021). https://doi.org/10.18196/jrc.2495

53. Chang, Y., Xu, J., Ghafoor, Z.: An IoT and Blockchain Approach for the Smart Water

Management System in Agriculture.

22,

105–116

(2021).

https://doi.org/10.12694:/scpe.v22i2.1869

54. Kumbi, A.A., Birje, M.N.: Deep CNN based Sunflower Atom Optimization Method for

Optimal Water Control in IoT.

Wirel Pers Commun.

122,

1221–1246

(2022).

https://doi.org/10.1007/s11277-021-08946-7

55. Torres, L., Romero, L., Aguirre, E., Ferro, R.: IoT Detection System for Mildew Disease in

Roses Using Neural Networks and Image Analysis. International Journal of Interactive Multimedia and Artificial Intelligence. 8, 105–116 (2023). https://doi.org/10.9781/ijimai.2023.07.001

56. Gagliardi, G., Lupia, M., Cario, G., Gaccio, F.C., D’angelo, V., Cosma, A.I.M., Casavola, A.:

An internet of things solution for smart agriculture.

Agronomy.

11,

(2021).

https://doi.org/10.3390/agronomy11112140

57. Salvini, G., Hofstede, G.J., Verdouw, C.N., Rijswijk, K., Klerkx, L.: Enhancing digital

transformation towards virtual supply chains: a simulation game for Dutch floriculture. Production Planning and Control. 33, 1252–1269 (2022). https://doi.org/10.1080/09537287.2020.1858361

58. Sangeetha, T., Periyathambi, E.: Automatic nutrient estimator: distributing nutrient solution in

hydroponic plants based on plant growth.

PeerJ Comput Sci.

10,

(2024).

https://doi.org/10.7717/peerj-cs.1871

59. Zhao, X., Feng, Y., Shi, X., Wang, Y., Zhang, G.: A color constancy based flower classification

method in the Blockchain data lake. Multimed Tools Appl. 83, 28657–28673 (2024). https://doi.org/10.1007/s11042-023-16656-4

60. Sumathi, M., Rajkamal, M., Raja, S.P., Venkatachalapathy, M., Vijayaraj, N.: A crop yield

prediction model based on an improved artificial neural network and yield monitoring using a Blockchain technique.

Int J Wavelets Multiresolut Inf Process.

20,

(2022).

https://doi.org/10.1142/S0219691322500308

61. Fadzly, N., Zuharah, W.F., Jenn Ney, J.W.: Can plants fool artificial intelligence? Using

machine learning to compare between bee orchids and bees. Plant Signal Behav. 16, (2021). https://doi.org/10.1080/15592324.2021.1935605

62. Jha, K., Doshi, A., Patel, P., Shah, M.: A comprehensive review on automation in agriculture

using artificial intelligence, (2019)

63. Wehrens, R., Afonso, M., Fonteijn, H., Paulo, J., Polder, G., Rijsbergen, M., van Hameren, G.,

Haegens, R., van den Helder, M., Zwinkels, H.: Determining flower colors from images using artificial intelligence. Euphytica. 220, (2024). https://doi.org/10.1007/s10681-023-03258-2

64. Herrera-Granda, I.D., Lorente-Leyva, L.L., Peluffo-Ordóñez, D.H., Alemany, M.M.E.: A

Forecasting Model to Predict the Demand of Roses in an Ecuadorian Small Business Under Uncertain Scenarios. (2020)

65. Zhang, Z., Rong, J., Qi, Z., Yang, Y., Zheng, X., Gao, J., Li, W., Yuan, T.: A multi-species pest

recognition and counting method based on a density map in the greenhouse. Comput Electron Agric. 217, (2024). https://doi.org/10.1016/j.compag.2023.108554

66. Kwon, H.S., Heo, S.: Predicting Vase Life of Cut Lisianthus Based on Biomass-Related

Characteristics Using AutoML.

Agriculture (Switzerland).

14,

(2024).

https://doi.org/10.3390/agriculture14091543

67. Afonso, M., Paulo, M.-J., Fonteijn, H., van den Helder, M., Zwinkels, H., Rijsbergen, M., van

Hameren, G., Haegens, R., Wehrens, R.: Automatic trait estimation in floriculture using computer vision and deep learning.

Smart Agricultural Technology.

7,

(2024).

https://doi.org/10.1016/j.atech.2023.100383

68. Sivaraj, M., Thanappan, R., Sharma, A.K.: A Comparative Analysis of AI Methods for Flower

Classification and Chemical Fingerprint Creation. International Research Journal of Multidisciplinary Technovation. 6, 241–260 (2024). https://doi.org/10.54392/irjmt24617

69. Caballero-Ramirez, D., Baez-Lopez, Y., Limon-Romero, J., Tortorella, G., Tlapa, D.: An

Assessment of Human Inspection and Deep Learning for Defect Identification in Floral Wreaths. Horticulture. 9, (2023). https://doi.org/10.3390/horticulturae9111213

70. Kwon, H.S., Heo, S.: Classification of Vase Life Day Based on Petal Colorimetric Data in Cut

Lisianthus Using AutoML.

Horticulture.

10,

(2024).

https://doi.org/10.3390/horticulturae10101033

71. Li, Y., Lv, Y., Ding, Y., Zhu, H., Gao, H., Zheng, L.: Research on a Flower Recognition Method

Based on Masked Autoencoders.

Horticulture.

10,

(2024).

https://doi.org/10.3390/horticulturae10050517

72. Viet, N.Q., Behdani, B., Bloemhof, J., Hoberg, K.: Value of data in multi-level supply chain

decisions: a case study in the Dutch floriculture sector. Int J Prod Res. 59, 1368–1385 (2021). https://doi.org/10.1080/00207543.2020.1821116

73. van Henten, E.J., Bac, C.W., Hemming, J., Edan, Y.: Robotics in protected cultivation. IFAC

Proceedings Volumes. 46, 170–177 (2013). https://doi.org/10.3182/20130828-2-SF-3019.00070

74. Pekkeriet, E.J., Van Henten, E.J.: Current developments of high-tech robotic and mechatronic

systems in horticulture and challenges for the future. (2011)

75. Priyadharshini, J., Arunn, T., Rithick Roshan, B., Vinoth Kumar, S.: Saffron Harvesting using

Robotic Arm Embedded on a Rover. In: 2023 2nd International Conference on Automation, Computing and Renewable Systems (ICACRS). pp. 1745–1750. IEEE (2023)

76. Patiluna, V.B., Maja, J.M.J., Robbins, J.: Evaluation of RFID power and UAV flight level in

plant inventory application. In: Bauer, C. and Thomasson, J.A. (eds.) Autonomous Air and Ground Sensing Systems for Agricultural Optimization and Phenotyping IX. p.7. SPIE (2024)

77. Ayaz, M., Ammad-Uddin, M., Sharif, Z., Mansour, A., Aggoune, E.-H.M.: Internet-of-Things

(IoT)-Based Smart Agriculture: Toward Making the Fields Talk. IEEE Access. 7, 129551–129583 (2019). https://doi.org/10.1109/ACCESS.2019.2932609

78. Rubanga, D.P., Hatanaka, K., Shimada, S.: Development of a simplified smart agriculture

system for small-scale greenhouse farming.

Sensors and Materials.

31,

(2019).

https://doi.org/10.18494/SAM.2019.2154

79. Zervopoulos, A., Tsipis, A., Alvanou, A.G., Bezas, K., Papamichail, A., Vergis, S., Stylidou,

A., Tsoumanis, G., Komianos, V., Koufoudakis, G., Koufoudakis, G., Oikonomou, K.: Wireless sensor network synchronization for precision agriculture applications. Agriculture (Switzerland). 10, (2020). https://doi.org/10.3390/agriculture10030089

80.

Eslamipoor, R., Sepehriyar, A.: Promoting green supply chain under carbon tax, carbon cap and carbon trading policies.

Bus Strategy Environ.

33,

4901–4912

(2024).

https://doi.org/10.1002/bse.3721

81. Van Henten, E.J.: Greenhouse mechanization: State of the art and future perspective. (2006)

82. Dusadeerungsikul, P.O., Nof, S.Y., Bechar, A., Tao, Y.: Collaborative Control Protocol for

Agricultural Cyber-physical System.

Procedia Manuf.

39,

235–242

(2019).

https://doi.org/10.1016/j.promfg.2020.01.330

83. Shree, C., Kaur, R., Upadhyay, S., Joshi, J.: Multi-feature Based Automated Flower Harvesting

Techniques in Deep Convolutional Neural Networking. In: Proceedings - 2019 4th International Conference on Internet of Things: Smart Innovation and Usages, IoT-SIU 2019 (2019)

84. Kumar, V.S., Gogul, I., Raj, M.D., Pragadesh, S.K., Sebastin, J.S.: Smart Autonomous

Gardening Rover with Plant Recognition Using Neural Networks. In: Procedia Computer Science. pp. 975–981 (2016)

85. Asefpour Vakilian, K., Massah, J.: A farmer-assistant robot for nitrogen fertilizing management

of greenhouse crops.

Comput Electron Agric.

139,

153–163

(2017).

https://doi.org/10.1016/j.compag.2017.05.012

References Chapter 3 Adamides, G. (2020). A review of climate-smart agriculture applications in Cyprus. Atmosphere, 11(9). https://doi.org/10.3390/ATMOS11090898 Afonso, M., Paulo, M.-J., Fonteijn, H., van den Helder, M., Zwinkels, H., Rijsbergen, M., van Hameren, G., Haegens, R., & Wehrens, R. (2024). Automatic trait estimation in floriculture using computer vision and deep learning. Smart Agricultural Technology, 7. https://doi.org/10.1016/j.atech.2023.100383 Ahmed, S., Nicholson, C. E., Rutter, S. R., Marshall, J. R., Perry, J. J., & Dean, J. R. (2023). Use of an unmanned aerial vehicle for monitoring and prediction of oilseed rape crop performance. PLOS ONE, 18(11), e0294184. https://doi.org/10.1371/journal.pone.0294184 Ahumada, O., & Villalobos, J. R. (2009). Application of planning models in the agri-food supply chain: A review. European Journal of Operational Research.

https://doi.org/10.1016/j.ejor.2008.02.014 Ahumada, O., & Villalobos, J. R. (2011). Operational model for planning the harvest and distribution of perishable agricultural products. International Journal of Production Economics, 133(2). https://doi.org/10.1016/j.ijpe.2011.05.015 Albornoz, V. M., Araneda, L. C., & Ortega, R. (2022). Planning and scheduling of selective harvest with management zones delineation. Annals of Operations Research, 316(2), 873–

890. https://doi.org/10.1007/s10479-021-04112-1

Albornoz, V. M., & Vera, P. I. (2023). Coordinating harvest planning and scheduling in an agricultural supply chain through a stochastic bilevel programming. International Transactions in Operational Research, 30(4), 1819–1842.

https://doi.org/https://doi.org/10.1111/itor.13190 Aliano Filho, A., A． Oliveira, W., & Melo, T. (2023). Multi-objective optimization for integrated sugarcane cultivation and harvesting planning. European Journal of Operational Research, 309(1), 330–344. https://doi.org/10.1016/j.ejor.2022.12.029 Aliano Filho, A., Cantane, D. R., Isler, P. R., & Oliveira Florentino, H. de. (2023). An integrated multi-objective mathematical model for sugarcane harvesting considering cumulative degree-days. Expert Systems with Applications, 232.

https://doi.org/10.1016/j.eswa.2023.120881 Asefpour Vakilian, K., & Massah, J. (2017). A farmer-assistant robot for nitrogen fertilizing management of greenhouse crops. Computers and Electronics in Agriculture, 139, 153–163. https://doi.org/10.1016/j.compag.2017.05.012 Atmaja, A. P., Hakim, A. E., Wibowo, A. P. A., & Pratama, L. A. (2021). Communication systems of smart agriculture based on wireless sensor networks in IoT. Journal of Robotics and Control (JRC), 2(4), 297–301. https://doi.org/10.18196/jrc.2495

Ayaz, M., Ammad-Uddin, M., Sharif, Z., Mansour, A., & Aggoune, E.-H. M. (2019). Internet-of- Things (IoT)-Based Smart Agriculture: Toward Making the Fields Talk. IEEE Access, 7, 129551–129583. https://doi.org/10.1109/ACCESS.2019.2932609 Babalola, A. O., Sundarakani, B., & Ganesh, K. (2011). Cold chain logistics in the floral industry. International Journal of Enterprise Network Management, 4(4), 400–413. https://doi.org/10.1504/IJENM.2011.043801 Boussios, D., Preckel, P. V, Yigezu, Y. A., Dixit, P. N., Akroush, S., M’hamed, H. C., Annabi, M., Aw‐Hassan, A., Shakatreh, Y., & Abdel Hadi, O. (2019). Modeling producer responses with dynamic programming: A case for adaptive crop management. Agricultural Economics, 50(1), 101–111.

Boussios, D., Preckel, P. V, Yigezu, Y. A., Dixit, P. N., Akroush, S., M’hamed, H. C., Annabi, M., Aw-Hassan, A., Shakatreh, Y., Abdel Hadi, O., Al-Abdallat, A., Abu El Enein, J., & Ayad, J. (2019). Modeling producer responses with dynamic programming: a case for adaptive crop management. Agricultural Economics (United Kingdom), 50(1), 101–111. https://doi.org/10.1111/agec.12469 Bumpus, W. (2013). NIST Cloud Computing Standards Roadmap. NIST Cloud Computing Standards.

Bunpalwong, M., Rukhiran, M., & Netinant, P. (2023). Improving marigold agriculture with an IoT-driven greenhouse irrigation management control system. Bulletin of Electrical Engineering and Informatics, 12(6), 3817–3825. https://doi.org/10.11591/eei.v12i6.6300 Burhan, I. M., Ali, Q. A., Hussein, I. S., & Jaleel, R. A. (2023). Mobile-computer Vision Model with Deep Learning for Testing Classification and Status of Flowers Images by using IoTs Devices. Journal of Wireless Mobile Networks, Ubiquitous Computing, and Dependable Applications, 14(1), 82–94. https://doi.org/10.58346/JOWUA.2023.I1.007 Caballero-Ramirez, D., Baez-Lopez, Y., Limon-Romero, J., Tortorella, G., & Tlapa, D. (2023). An Assessment of Human Inspection and Deep Learning for Defect Identification in Floral Wreaths. Horticulturae, 9(11). https://doi.org/10.3390/horticulturae9111213 Caicedo Solano, N. E., García Llinás, G. A., Montoya-Torres, J. R., & Ramirez Polo, L. E. (2020). A planning model of crop maintenance operations inspired in lean manufacturing. Computers and Electronics in Agriculture, 179.

https://doi.org/10.1016/j.compag.2020.105852 Caixeta-Filho, J. V., van Swaay-Neto, J. M., De Pádua Wagemaker, A., & Wagemaker, A. (2002). Optimization of the Production Planning and Trade of Lily Flowers at Jan de Wit Company. Interfaces, 32(1), 35–46. https://doi.org/10.1287/inte.32.1.35.13

Caixeta-Filho, J. V., van Swaay-Neto, J. M., & Lopes, R. L. (2000). Linear programming applied to the flower sector: a Gladiolus bulb production case study. International Transactions in Operational Research, 7, 525–537. www.elsevier.com/locate/orms Cedillo Villavicencio, C. J., González Carrión, C. M., Salcedo-Muñoz, V. E., & Sotomayor- Pereira, J. G. (2021). El sector florícola del Ecuador y su aporte a la balanza comercial agropecuaria: periodo 2009 – 2020. Revista Científica y Tecnológica UPSE, 8(1), 74–82. https://doi.org/10.26423/rctu.v8i1.549 Chan, A., Cruz, D., & Sy, C. (2025). A multi-objective optimization approach for the design of circular supply chains of perishable agricultural products. Cleaner Waste Systems, 12. https://doi.org/10.1016/j.clwas.2025.100347 Chang, Y., Xu, J., & Ghafoor, Z. (2021). An IoT and Blockchain Approach for the Smart Water Management System in Agriculture. 22(2), 105–116.

https://doi.org/10.12694:/scpe.v22i2.1869 Chen, H., & Ryan, J. K. (2023). Optimal specialty crop planning policies with yield learning and forward contract. Production and Operations Management, 32(2), 359–378. https://doi.org/10.1111/POMS.13842 Custodio, J. M., Billones, R. K., Concepcion, R., & Vicerra, R. R. (2024). Optimization of crop harvesting schedules and land allocation through linear programming. Process Integration and Optimization for Sustainability, 8(1), 123–134.

de Carvalho, P. R. V., Naoum-Sawaya, J., & Elhedhli, S. (2022). Blockchain-Enabled supply chains: An application in fresh-cut flowers. Applied Mathematical Modelling, 110, 841–858. https://doi.org/10.1016/j.apm.2022.06.011 Dole, J. M. ., & Wilkins, H. F. . (2005). Floriculture : principles and species. 1023. https://search.worldcat.org/title/53887461 Drechsler, M., & Holzapfel, A. (2022). Decision Support in Horticultural Supply Chains: A Planning Problem Framework for Small and Medium-Sized Enterprises. Agriculture (Switzerland), 12(11). https://doi.org/10.3390/agriculture12111922 Dusadeerungsikul, P. O., Nof, S. Y., Bechar, A., & Tao, Y. (2019). Collaborative Control Protocol for Agricultural Cyber-physical System. Procedia Manufacturing, 39, 235–242. https://doi.org/10.1016/j.promfg.2020.01.330 El-Telbany, O., Abdelghaffar, H., & Amin, H. (2020, June). Exploring the digital transformation gap: Evidence from organizations in emerging economies. Proceedings of the 24th Pacific Asia Conference on Information Systems: Information Systems (IS) for the Future, PACIS

2020.

Escallón-Barrios, M., Castillo-Gomez, D., Leal, J., Montenegro, C., & Medaglia, A. L. (2022). Improving harvesting operations in an oil palm plantation. Annals of Operations Research, 314(2), 411–449. https://doi.org/10.1007/s10479-020-03686-6 Eslamipoor, R., & Sepehriyar, A. (2024). Promoting green supply chain under carbon tax, carbon cap and carbon trading policies. Business Strategy and the Environment, 33(5), 4901–4912. https://doi.org/10.1002/bse.3721 Estrada, J. S., Vasconez, J. P., Fu, L., & Cheein, F. A. (2024). Deep Learning based flower detection and counting in highly populated images: A peach grove case study. Journal of Agriculture and Food Research, 15. https://doi.org/10.1016/j.jafr.2023.100930 Et-taibi, B., Abid, M. R., Boufounas, E.-M., Morchid, A., Bourhnane, S., Abu Hamed, T., & Benhaddou, D. (2024). Enhancing water management in smart agriculture: A cloud and IoT- Based smart irrigation system. Results in Engineering, 22, 102283. https://doi.org/10.1016/j.rineng.2024.102283 Expoflores. (2024). Reporte estadístico anual. www.expoflores.com Fadzly, N., Zuharah, W. F., & Jenn Ney, J. W. (2021). Can plants fool artificial intelligence? Using machine learning to compare between bee orchids and bees. Plant Signaling and Behavior, 16(10). https://doi.org/10.1080/15592324.2021.1935605 Fathollahi-Fard, A. M., Tian, G., Ke, H., Fu, Y., & Wong, K. Y. (2023). Efficient multi-objective metaheuristic algorithm for sustainable harvest planning problem. Computers and Operations Research, 158. https://doi.org/10.1016/j.cor.2023.106304 Faust, J. E., & Dole, J. M. (2021). The global cut flower and foliage marketplace. Cut Flowers and Foliages, 27, 1–47.

Faust, J. E., Dole, J. M., & Lopez, R. G. (2016). The floriculture vegetative cutting industry. Horticultural Reviews, 44, 121–172.

Ferrer, J. C., Mac Cawley, A., Maturana, S., Toloza, S., & Vera, J. (2008). An optimization approach for scheduling wine grape harvest operations. International Journal of Production Economics, 112(2), 985–999. https://doi.org/10.1016/j.ijpe.2007.05.020 Fiorucci, A. S., & Fankhauser, C. (2017). Plant Strategies for Enhancing Access to Sunlight. In Current Biology (Vol. 27, Number 17, pp. R931–R940). Cell Press. https://doi.org/10.1016/j.cub.2017.05.085 Gagliardi, G., Lupia, M., Cario, G., Gaccio, F. C., D’angelo, V., Cosma, A. I. M., & Casavola, A. (2021). An internet of things solution for smart agriculture. Agronomy, 11(11). https://doi.org/10.3390/agronomy11112140 Gómez-Lagos, J. E., González-Araya, M. C., Soto-Silva, W. E., & Rivera-Moraga, M. M. (2021). Optimizing tactical harvest planning for multiple fruit orchards using a

metaheuristic modeling approach. European Journal of Operational Research, 290(1), 297–

312. https://doi.org/10.1016/j.ejor.2020.08.015

Grand View Research. (2024). Cut Flowers Market Analysis 2018-2030. Guaita-Pradas, I., Rodríguez-Mañay, L. O., & Marques-Perez, I. (2023). Competitiveness of Ecuador’s flower industry in the global market in the period 2016–2020. Sustainability,

15(7), 5821.

Hartmann, F., Van Heck, E., Murray, C., Van Driel, H., Ahmedi, S., Wierenga, B., Althuizen, N., Chen, B., & Yuan, Y. (2017). Selling flowers with analytics. www.rsm.nl/discovery Hassan, S. I., Alam, M. M., Illahi, U., Al Ghamdi, M. A., Almotiri, S. H., & Su’ud, M. M. (2021). A Systematic Review on Monitoring and Advanced Control Strategies in Smart Agriculture. IEEE Access, 9, 32517–32548. https://doi.org/10.1109/ACCESS.2021.3057865 Hatfield, J. L., & Prueger, J. H. (2015). Temperature extremes: Effect on plant growth and development. Weather and Climate Extremes, 10, 4–10.

https://doi.org/10.1016/j.wace.2015.08.001 Heemskerk, C., Boode, A. H., Arntzen, P., & Fesselet, L. (2020). HiPerGreen: Greenhouse crop scouting by a UAS: Lessons learned from cross-domain cooperation in applied research.

2020 21st International Conference on Research and Education in Mechatronics, REM

2020. https://doi.org/10.1109/REM49740.2020.9313899

Herrera-Cáceres, C., Pérez-Galarce, F., Álvarez-Miranda, E., & Candia-Véjar, A. (2017). Optimization of the harvest planning in the olive oil production: A case study in Chile. Computers and Electronics in Agriculture, 141, 147–159.

https://doi.org/10.1016/j.compag.2017.07.017 Herrera-Granda, I. D., Lorente-Leyva, L. L., Peluffo-Ordóñez, D. H., & Alemany, M. M. E. (2020). A Forecasting Model to Predict the Demand of Roses in an Ecuadorian Small Business Under Uncertain Scenarios. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics):

12566 LNCS. https://doi.org/10.1007/978-3-030-64580-9_21

Higaki, T., Rasmussen, H. P., & Carpenter, W. J. (1984). A study of Some Morphological and Anatomical Aspects of Anthurium andreanum Lind. University of Hawaii. http://hdl.handle.net/10125/25543 Hoogeveen, H., Tomczyk, J., & van der Zanden, T. C. (2019). Flower power: Finding optimal plant cutting strategies through a combination of optimization and data mining. Computers & Industrial Engineering, 127, 39–44.

Hosny, K. M., El-Hady, W. M., & Samy, F. M. (2024). Technologies, Protocols, and applications of Internet of Things in greenhouse Farming: A survey of recent advances. In Information

Processing in Agriculture. China Agricultural University.

https://doi.org/10.1016/j.inpa.2024.04.002 Jarumaneeroj, P., Laosareewatthanakul, N., & Akkerman, R. (2021). A multi-objective approach to sugarcane harvest planning in Thailand: Balancing output maximization, grower equity, and supply chain efficiency. Computers and Industrial Engineering, 154. https://doi.org/10.1016/j.cie.2021.107129 Jawaharlal, M., Ganga, M., Padmadevi, K., Jegadeeswari, V., & Karthikeyan, S. (2010). A technical guide on carnation.

Jha, K., Doshi, A., Patel, P., & Shah, M. (2019). A comprehensive review on automation in agriculture using artificial intelligence. In Artificial Intelligence in Agriculture (Vol. 2, pp. 1–12). KeAi Communications Co. https://doi.org/10.1016/j.aiia.2019.05.004 Junqueira, R. de Á. R., & Morabito, R. (2019). Modeling and solving a sugarcane harvest front scheduling problem. International Journal of Production Economics, 213, 150–160. https://doi.org/10.1016/j.ijpe.2019.03.009 Juroszek, P., & Von Tiedemann, A. (2013). Plant pathogens, insect pests and weeds in a changing global climate: A review of approaches, challenges, research gaps, key studies and concepts. Journal of Agricultural Science, 151(2), 163–188.

https://doi.org/10.1017/S0021859612000500 Kamaruddin, F., Malik, N. N. N. A., Murad, N. A., Latiff, N. M. A., Yusof, S. K. S., & Hamzah, S. A. (2019). IoT-based intelligent irrigation management and monitoring system using arduino. Telkomnika (Telecommunication Computing Electronics and Control), 17(5), 2378–2388. https://doi.org/10.12928/TELKOMNIKA.v17i5.12818 Kamyab, H., Khademi, T., Chelliapan, S., SaberiKamarposhti, M., Rezania, S., Yusuf, M., Farajnezhad, M., Abbas, M., Hun Jeon, B., & Ahn, Y. (2023). The latest innovative avenues for the utilization of artificial Intelligence and big data analytics in water resource management. Results in Engineering, 20, 101566.

https://doi.org/10.1016/j.rineng.2023.101566 Khalilzadeh, Z., & Wang, L. (2022). Corn planting and harvest scheduling under storage capacity and growing degree units uncertainty. Scientific Reports, 12(1). https://doi.org/10.1038/s41598-022-25797-9 Kumar, V. S., Gogul, I., Raj, M. D., Pragadesh, S. K., & Sebastin, J. S. (2016). Smart Autonomous Gardening Rover with Plant Recognition Using Neural Networks. Procedia Computer Science, 93, 975–981. https://doi.org/10.1016/j.procs.2016.07.289 Kumbi, A. A., & Birje, M. N. (2022). Deep CNN based Sunflower Atom Optimization Method for Optimal Water Control in IoT. Wireless Personal Communications, 122(2), 1221–1246. https://doi.org/10.1007/s11277-021-08946-7

Kusumastuti, R. D., Donk, D. P. Van, & Teunter, R. (2016). Crop-related harvesting and processing planning: A review. International Journal of Production Economics, 174, 76–92. https://doi.org/10.1016/j.ijpe.2016.01.010 Kwon, H. S., & Heo, S. (2024a). Classification of Vase Life Day Based on Petal Colorimetric Data in Cut Lisianthus Using AutoML. Horticulturae, 10(10).

https://doi.org/10.3390/horticulturae10101033 Kwon, H. S., & Heo, S. (2024b). Predicting Vase Life of Cut Lisianthus Based on Biomass- Related Characteristics Using AutoML. Agriculture (Switzerland), 14(9). https://doi.org/10.3390/agriculture14091543 Lezoche, M., Panetto, H., Kacprzyk, J., Hernandez, J. E., & Alemany Díaz, M. M. E. (2020). Agri-food 4.0: A survey of the Supply Chains and Technologies for the Future Agriculture. Computers in Industry, 117, 103187. https://doi.org/10.1016/j.compind.2020.103187 Li, Y., Lv, Y., Ding, Y., Zhu, H., Gao, H., & Zheng, L. (2024). Research on a Flower Recognition Method Based on Masked Autoencoders. Horticulturae, 10(5).

https://doi.org/10.3390/horticulturae10050517 Liu, Y., Ma, X., Shu, L., Hancke, G. P., & Abu-Mahfouz, A. M. (2021). From Industry 4.0 to Agriculture 4.0: Current Status, Enabling Technologies, and Research Challenges. IEEE Transactions on Industrial Informatics, 17(6), 4322–4334.

https://doi.org/10.1109/TII.2020.3003910 López, M. M., Vera Andreo, J., Plà Aragonés, L. M., & Recalde-Ramírez, J. L. (2024). Design of a mathematical model to optimize farmer food security and promote rural development in Paraguay. Annals of Operations Research. https://doi.org/10.1007/s10479-024-06199-8 Mahdavimanshadi, M., Yao, S., & Fan, N. (2024). Optimal guayule harvest planning and machinery scheduling under drought scenarios in semi-arid farms. Smart Agricultural Technology, 7. https://doi.org/10.1016/j.atech.2024.100420 Mana, A. A., Allouhi, A., Hamrani, A., Rahman, S., el Jamaoui, I., & Jayachandran, K. (2024). Sustainable AI-based production agriculture: Exploring AI applications and implications in agricultural practices. Smart Agricultural Technology, 7.

https://doi.org/10.1016/j.atech.2024.100416 Mann, H. M. R., Iosifidis, A., Jepsen, J. U., Welker, J. M., Loonen, M. J. J. E., & Høye, T. T. (2022). Automatic flower detection and phenology monitoring using time-lapse cameras and deep learning. Remote Sensing in Ecology and Conservation, 8(6), 765–777. https://doi.org/10.1002/rse2.275 Mantilla, F., Mejía, G., & Tascón, D. (2025). The role of industry 4.0 technologies in the export flower industry: Insights from a systematic literature review and surveys in emerging

economies. In Results in Engineering (Vol. 25). Elsevier B.V. https://doi.org/10.1016/j.rineng.2025.104507 Market Analysis Report. (2023). Floriculture Market-Global Industry Analysis and Forecast. Mejía, G., Guerrero, W., Sarmiento, A., Serrano, N., Sarmiento, M., & Sánchez, C. (2019). Delivery of Perishable Export Products in Smart Cities: A Case Study in Bogotá (Colombia). Procedia Manufacturing, 39(2019), 1946–1952.

https://doi.org/10.1016/j.promfg.2020.01.237 Mejía, G., Sarmiento, A., & Guerrero, W. (2022). Truck and Cross-Docking Scheduling in the Flower Export Sector in Colombia. Transportation Research Record, 03611981221130332. Monroy, N., Cure, J. R., Pérez, I., & Cure, J. R. (2014). Estudio de la variabilidad en el clima y la producción de rosas en la sabana de Bogotá. Revista de Ingeniería, 14, 38–43. Montenegro-Dos Santos, F., Pérez-Galarce, F., Monardes-Concha, C., Candia-Véjar, A., Seido- Nagano, M., & Gómez-Lagos, J. (2023). A Rolling Horizon scheme for rescheduling in agricultural harvest. Computers and Electronics in Agriculture, 215. https://doi.org/10.1016/j.compag.2023.108392 Morchid, A., Jebabra, R., Khalid, H. M., El Alami, R., Qjidaa, H., & Ouazzani Jamil, M. (2024). IoT-based smart irrigation management system to enhance agricultural water security using embedded systems, telemetry data, and cloud computing. Results in Engineering, 23,

102829. https://doi.org/10.1016/j.rineng.2024.102829

Morchid, A., Oughannou, Z., Alami, R. El, Qjidaa, H., Jamil, M. O., & Khalid, H. M. (2024). Integrated internet of things (IoT) solutions for early fire detection in smart agriculture. Results in Engineering, 24, 103392. https://doi.org/10.1016/j.rineng.2024.103392 Nagasawa, H., Kotani, M., & Morizawa, K. (2009). Optimal Cooperative Harvesting Patterns of Agricultural Fresh Products in Case of Multiple Farmers and Multiple Markets under Periodical Flowering. Journal of the Operations Research Society of Japan, 52(4), 417–432. https://doi.org/10.15807/jorsj.52.417 Navarrete, D. P., Ecuador, G., Knorringa, P., & Gomez, G. (2018). Cut-Roses Global Value Chain Governance: Ecuadorian Exports to the Netherlands.

P. Senna, P., Barros, A. C., Bonnin Roca, J., & Azevedo, A. (2023). Development of a digital maturity model for Industry 4.0 based on the technology-organization-environment framework. Computers & Industrial Engineering, 185, 109645.

https://doi.org/10.1016/J.CIE.2023.109645 Park, A., & Li, H. (2021). The effect of Blockchain technology on supply chain sustainability performances. Sustainability (Switzerland), 13(4), 1–18.

https://doi.org/10.3390/SU13041726

Patel, D. H., Shah, K. P., Gupta, R., Jadav, N. K., Tanwar, S., Neagu, B. C., Attila, S., Alqahtani, F., & Tolba, A. (2023). Blockchain-Based Crop Recommendation System for Precision Farming in IoT Environment. Agronomy, 13(10).

https://doi.org/10.3390/agronomy13102642 Patiluna, V. B., Maja, J. M. J., & Robbins, J. (2024). Evaluation of RFID power and UAV flight level in plant inventory application. In C. Bauer & J. A. Thomasson (Eds.), Autonomous Air and Ground Sensing Systems for Agricultural Optimization and Phenotyping IX (Vol. 13053, p. 7). SPIE. https://doi.org/10.1117/12.3021150 Pekkeriet, E. J., & Van Henten, E. J. (2011). Current developments of high-tech robotic and mechatronic systems in horticulture and challenges for the future. In Acta Horticulturae (Vol. 893). https://doi.org/10.17660/ActaHortic.2011.893.4 Pekkerieta, E. J., Van Henten, E. J., & Campen, J. B. (2015). Contribution of innovative technologies to new developments in horticulture. In Acta Horticulturae (Vol. 1099). Pérez, I., Cure, J. R., & Monroy, N. (2002). Modelo de predicción y manejo de cultivos de rosas. Revista de Ingeniería, 15, 18.

Priyadharshini, J., Arunn, T., Rithick Roshan, B., & Vinoth Kumar, S. (2023). Saffron Harvesting using Robotic Arm Embedded on a Rover. 2023 2nd International Conference on Automation, Computing and Renewable Systems (ICACRS), 1745–1750. https://doi.org/10.1109/ICACRS58579.2023.10404995 Reid, M. S., & Jiang, C.-Z. (2012). Postharvest Biology and Technology of Cut Flowers and Potted Plants. Horticultural Reviews, 40.

Rocco, C. D., & Morabito, R. (2016). Production and logistics planning in the tomato processing industry: A conceptual scheme and mathematical model. Computers and Electronics in Agriculture, 127, 763–774. https://doi.org/10.1016/j.compag.2016.08.002 Rodríguez, S., Gualotuña, T., & Grilo, C. (2017). A System for the Monitoring and Predicting of Data in Precision Agriculture in a Rose Greenhouse Based on Wireless Sensor Networks. Procedia Computer Science, 121, 306–313. https://doi.org/10.1016/j.procs.2017.11.042 Rosero-Montalvo, P. D., Erazo-Chamorro, V. C., López-Batista, V. F., Moreno-García, M. N., & Peluffo-Ordóñez, D. H. (2020). Environment monitoring of rose crops greenhouse based on autonomous vehicles with a wsn and data analysis. Sensors (Switzerland), 20(20), 1–15. https://doi.org/10.3390/s20205905 Roumeliotis, C., Dasygenis, M., Lazaridis, V., & Dossis, M. (2024). Blockchain and Digital Twins in Smart Industry 4.0: The Use Case of Supply Chain-A Review of Integration Techniques and Applications. Designs, 8(6), 105. https://doi.org/10.3390/designs8060105

Rubanga, D. P., Hatanaka, K., & Shimada, S. (2019). Development of a simplified smart agriculture system for small-scale greenhouse farming. Sensors and Materials, 31(3). https://doi.org/10.18494/SAM.2019.2154 Ruiz-Torres, A. J., Villalobos, J. R., Salvador-Jijon, M., & Alomoto, N. (2012). Planning models for floriculture operations. International Journal of Applied Management Science, 4(2), 148–164.

Salvini, G., Hofstede, G. J., Verdouw, C. N., Rijswijk, K., & Klerkx, L. (2022). Enhancing digital transformation towards virtual supply chains: a simulation game for Dutch floriculture. Production Planning and Control, 33(13), 1252–1269.

https://doi.org/10.1080/09537287.2020.1858361 Sangeetha, T., & Periyathambi, E. (2024). Automatic nutrient estimator: distributing nutrient solution in hydroponic plants based on plant growth. PeerJ Computer Science, 10. https://doi.org/10.7717/peerj-cs.1871 Sauer, P. C., & Seuring, S. (2023). How to conduct systematic literature reviews in management research: a guide in 6 steps and 14 decisions. Review of Managerial Science, 17(5), 1899–1933. https://doi.org/10.1007/s11846-023-00668-3 Sekaran, K., Meqdad, M. N., Kumar, P., Rajan, S., & Kadry, S. (2020). Smart agriculture management system using internet of things. Telkomnika (Telecommunication Computing Electronics and Control), 18(3), 1275–1284.

https://doi.org/10.12928/TELKOMNIKA.v18i3.14029 Sepehriar, A., & Eslamipoor, R. (2024). An economical single-vendor single-buyer framework for carbon emission policies. Journal of Business Economics, 94(6), 927–945. https://doi.org/10.1007/s11573-024-01192-w Shree, C., Kaur, R., Upadhyay, S., & Joshi, J. (2019). Multi-feature Based Automated Flower Harvesting Techniques in Deep Convolutional Neural Networking. Proceedings - 2019 4th International Conference on Internet of Things: Smart Innovation and Usages, IoT-SIU

2019. https://doi.org/10.1109/IoT-SIU.2019.8777338

Sivaraj, M., Thanappan, R., & Sharma, A. K. (2024). A Comparative Analysis of AI Methods for Flower Classification and Chemical Fingerprint Creation. International Research Journal of Multidisciplinary Technovation, 6(6), 241–260. https://doi.org/10.54392/irjmt24617 Spherical Insights LLP. (2023). Global Cut Flowers Market Size. Sumathi, M., Rajkamal, M., Raja, S. P., Venkatachalapathy, M., & Vijayaraj, N. (2022). A crop yield prediction model based on an improved artificial neural network and yield monitoring using a Blockchain technique. International Journal of Wavelets, Multiresolution and Information Processing, 20(6). https://doi.org/10.1142/S0219691322500308

Sun, X., Li, Z., Zhu, T., & Ni, C. (2021). Four-dimension deep learning method for flower quality grading with depth information. Electronics (Switzerland), 10(19). https://doi.org/10.3390/electronics10192353 Taşkıner, T., & Bilgen, B. (2021). Optimization Models for Harvest and Production Planning in Agri-Food Supply Chain: A Systematic Review. In Logistics (Vol. 5, Number 3). MDPI. https://doi.org/10.3390/logistics5030052 Tornatzky, G. , L., & Fleischer, M. (1990). The Processes of Technological Innovation, Lexington, MA: Lexington Books, 1990. In Technological Innovation (Number January

1990).

Torres, L., Romero, L., Aguirre, E., & Ferro, R. (2023). IoT Detection System for Mildew Disease in Roses Using Neural Networks and Image Analysis. International Journal of Interactive Multimedia and Artificial Intelligence, 8(4), 105–116. https://doi.org/10.9781/ijimai.2023.07.001 Tsai, W.-C., & Shen, C.-W. (2024). Using a smart contract for the floral supply chain. Asia Pacific Management Review. https://doi.org/10.1016/j.apmrv.2023.12.004 van Henten, E. J. (2006). Greenhouse mechanization: state of the art and future perspective. Acta Horticulturae, 710(710), 55–70. https://doi.org/10.17660/ActaHortic.2006.710.3 Van Henten, E. J. (2006). Greenhouse mechanization: State of the art and future perspective. In Acta Horticulturae (Vol. 710). https://doi.org/10.17660/actahortic.2006.710.3 van Henten, E. J., Bac, C. W., Hemming, J., & Edan, Y. (2013). Robotics in protected cultivation. IFAC Proceedings Volumes, 46(18), 170–177. https://doi.org/10.3182/20130828-2-SF-

3019.00070

Van Hentenryck, P., & Dalmeijer, K. (2024). AI4OPT: AI Institute for Advances in Optimization. AI Magazine, 45(1), 42–47. https://doi.org/10.1002/aaai.12146 Vanegas López, J. G., Merlos García, J. J., & Mayorga Abril, C. M. (2017). Flower Export Barriers: A Comparative Study in Colombia, Mexico and Ecuador. Latin American Business Review, 18(3–4), 227–250. https://doi.org/10.1080/10978526.2017.1354705 Varas, M., Basso, F., Maturana, S., Osorio, D., & Pezoa, R. (2020). A multi-objective approach for supporting wine grape harvest operations. Computers and Industrial Engineering, 145. https://doi.org/10.1016/j.cie.2020.106497 Vasco, C., Ortega, C., Cepeda, D., Salazar, D., & Pazmiño, J. (2025). The socioeconomic drivers of pesticide use in floriculture: insights from greenhouse rose production in Ecuador. Journal of Environmental Studies and Sciences, 15(3), 550–560. https://doi.org/10.1007/s13412-024-00964-8

Viet, N. Q., Behdani, B., Bloemhof, J., & Hoberg, K. (2021). Value of data in multi-level supply chain decisions: a case study in the Dutch floriculture sector. International Journal of Production Research, 59(5), 1368–1385. https://doi.org/10.1080/00207543.2020.1821116 Villalobos, J. R., Salvador-Jijon, M., & Alomoto, N. (2012). Planning models for floriculture operations. In Int. J. Applied Management Science (Vol. 4, Number 2). Wang, S., Liu, H., Qin, X., Dai, J., & Liu, J. (2023). Cherry blossom and ginkgo leaf coloration phenology dataset of China from 2009 to 2019 extracted from big data. Geoscience Data Journal. https://doi.org/10.1002/gdj3.231 Wang, X. (Annie), Tang, J., & Whitty, M. (2021). DeepPhenology: Estimation of apple flower phenology distributions based on deep learning. Computers and Electronics in Agriculture,

185. https://doi.org/10.1016/j.compag.2021.106123

Wani, M. A., Din, A., Nazki, I. T., Rehman, T. U., Al-Khayri, J. M., Jain, S. M., Lone, R. A., Bhat, Z. A., & Mushtaq, M. (2023). Navigating the future: exploring technological advancements and emerging trends in the sustainable ornamental industry. Frontiers in Environmental Science, 11, 1188643. https://doi.org/10.3389/fenvs.2023.1188643 Wehrens, R., Afonso, M., Fonteijn, H., Paulo, J., Polder, G., Rijsbergen, M., van Hameren, G., Haegens, R., van den Helder, M., & Zwinkels, H. (2024). Determining flower colors from images using artificial intelligence. Euphytica, 220(1). https://doi.org/10.1007/s10681-023-

03258-2

What Is Artificial Intelligence (AI)? | IBM. (n.d.). Retrieved December 16, 2024, from https://www.ibm.com/think/topics/artificial-intelligence Widodo, K. H., Nagasawa, H., Morizawa, K., & Ota, M. (2006). A periodical floweringharvesting model for delivering agricultural fresh products. European Journal of Operational Research, 170(1), 24–43. https://doi.org/10.1016/j.ejor.2004.05.024 Yang, J., Sharma, A., & Kumar, R. (2021). IoT-based framework for smart agriculture. International Journal of Agricultural and Environmental Information Systems, 12(2), 1–14. https://doi.org/10.4018/IJAEIS.20210401.oa1 Yao, S., Fan, N., Seavert, C., & Teegerstrom, T. (2023). Demand-Driven Harvest Planning and Machinery Scheduling for Guayule. Operations Research Forum, 4(1). https://doi.org/10.1007/s43069-022-00192-2 Zervopoulos, A., Tsipis, A., Alvanou, A. G., Bezas, K., Papamichail, A., Vergis, S., Stylidou, A., Tsoumanis, G., Komianos, V., Koufoudakis, G., Koufoudakis, G., & Oikonomou, K. (2020). Wireless sensor network synchronization for precision agriculture applications. Agriculture (Switzerland), 10(3). https://doi.org/10.3390/agriculture10030089 Zhang, T., Vail, S., Duddu, H. S. N., Parkin, I. A. P., Guo, X., Johnson, E. N., & Shirtliffe, S. J. (2021). Phenotyping Flowering in Canola (Brassica napus L.) and Estimating Seed Yield

Using an Unmanned Aerial Vehicle-Based Imagery. Frontiers in Plant Science, 12, 686332. https://doi.org/10.3389/fpls.2021.686332 Zhang, Y., & Swaminathan, J. M. (2020). Improved crop productivity through optimized planting schedules. Manufacturing & Service Operations Management, 22(6), 1165–1180. Zhang, Z., Rong, J., Qi, Z., Yang, Y., Zheng, X., Gao, J., Li, W., & Yuan, T. (2024). A multispecies pest recognition and counting method based on a density map in the greenhouse. Computers and Electronics in Agriculture, 217.

https://doi.org/10.1016/j.compag.2023.108554 Zhao, R., Wang, L., Zhang, X., Zhang, Y., Wang, L., & Peng, H. (2018). A OneM2M-Compliant Stacked Middleware Promoting IoT Research and Development. IEEE Access, 6, 63546–

63559. https://doi.org/10.1109/ACCESS.2018.2876197

Zhao, X., Feng, Y., Shi, X., Wang, Y., & Zhang, G. (2024). A color constancy based flower classification method in the Blockchain data lake. Multimedia Tools and Applications, 83(10), 28657–28673. https://doi.org/10.1007/s11042-023-16656-4 Zuniga Vazquez, D. A., Fan, N., Teegerstrom, T., Seavert, C., Summers, H. M., Sproul, E., & Quinn, J. C. (2021). Optimal production planning and machinery scheduling for semi-arid farms. Computers and Electronics in Agriculture, 187.

https://doi.org/10.1016/j.compag.2021.106288