An ontology - driven conceptual modeling of olive oil supply chain

Autor: Ailí Andrea Tamayo López Tutor: Djamel Rahmani External tutor: Giuliano Vitali, University of Bologna 12 / June / 2023 Bachelors degree thesis Agricultural Science Engineering AN ONTOLOGY - DRIVEN CONCEPTUAL MODELING OF OLIVE OIL SUPPLY CHAIN An ontology - driven conceptual modeling of olive oil supply chain 1 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Abstract The agri-food sector is one of the largest and most important sectors of today's industrial economy. Production and economic losses in the food market and along the food production chain are one of the main problems faced by this industry, mainly due to poor coordination among the different actors involved in a supply chain (SC). In SC, a main obstacle to increasing efficiency is the lack of an integrated flow of information data between the different suppliers and buyers in the sector. Partners in the agri-food sector SC have different experiences and roles, resulting in language asymmetry that affects overall production performance due to information leakage in some parts of the chain. Communication is critical not only from person to person, but also from person to machine or from machine to machine. Industry 4.0 and IoT now permeate most scenarios, from the field to the fork, and the scene is also populated by ICT engineers who use a very specialized language and are involved in the development of analytics tools. These kinds of tools (e.g., computational models, life-cycle assessment) promise better performance that increases productivity and reduces losses, they introduce a different language. For the reasons outlined, a common language needs to be defined, and one of the solutions studied in recent years to improve semantic communication is represented by ontologies, which identify and describe a specific vocabulary by domain, consisting of annotated terms with a specific definition and related to each other. Ontologies help to link concepts and processes in the industry, contributing to the understanding and improvement of the complex system including markets and supply chains. They already represent a way to standardize communication in the agri-food industry and have also been studied in recent years. This thesis begins with an introduction to a general analysis of supply chains, together with the relationship with value chains alongside their different objectives. Then, a study of the structure and functionality of ontologies based on the literature is addressed with a description of previous ontologies and the tools most useful for working with them. Finally, an attempt is made to create a new ontology for the field of olive oil SC, a discussion is made about its potential usefulness and final conclusions are drawn. An ontology - driven conceptual modeling of olive oil supply chain 2 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Resum El sector agroalimentari és un dels sectors més grans i importants de l'economia industrial actual. Les pèrdues de producció i econòmiques al mercat alimentari i al llarg de la cadena de producció alimentària són un dels principals problemes als quals s'enfronta aquesta indústria, principalment a causa de la manca de coordinació entre els diferents actors implicats a la cadena de producció (CP). En una CP, un dels principals obstacles per augmentar l'eficiència és la manca d'un flux integrat d'informació entre els diferents proveïdors i compradors del sector. Els socis de la CP del sector agroalimentari tenen diferents experiències i rols, el que resulta en una asimetría lingüística que afecta el rendiment de producció global a causa de pèrdua d'informació en algunes parts de la cadena. La comunicació és crucial no només de persona a persona, sinó també de persona a màquina o de màquina a màquina. La Indústria 4.0 i l'Internet de les Coses (IoT) impregnen ara la majoria del desenvolupament de la CP, des del camp fins a la taula, trobant nos també l’introducció d’enginyers de TIC que introdueixen en la industria un llenguatge molt especialitzat per tal d'aplicar el desenvolupament d'eines d'anàlisi. Aquestes eines (per exemple, models computacionals, avaluació del cicle de vida) prometen un millor rendiment que ajuda a augmenta la productivitat i redueix les pèrdues, introduint un llenguatge diferent. Per les raons esmentades, cal definir un llenguatge comú, i una de les solucions estudiades en els últims anys per millorar la comunicació semàntica està representada per les ontologies. Aquestes identifiquen i descriuen un vocabulari específic per a un domini, format per termes anotats amb una definició específica i relacionats entre ells. Les ontologies ajuden a vincular conceptes i processos a la indústria, contribuint a la comprensió i millora dels sistemes complexos, incloent mercats i cadenes de subministrament. El estudi de les ontologies s’ha estudiat molt en els últims anys ja comencen a agafar importància per representar una manera d'estandaritzar la comunicació en la indústria agroalimentària. Aquest treball de grau comença amb una introducció a una anàlisi general de les cadenes de producció, juntament amb la relació a les cadenes de valor i els seus diferents objectius. A continuació, s'aborda un estudi de l'estructura i funcionalitat de les ontologies basat en literatura que ens permet definir que son les ontologies com i en quins àmbits de la industria agroalimentaria s’estan utilitzant i quines son les eines més útilitzades per treballar amb elles. Finalment, es fa un intent de creació d’una nova ontologia per a la cadena de producció de l'oli d'oliva, la qual s'avalua la seva utilitat i aplicacions en el punt de discussió. An ontology - driven conceptual modeling of olive oil supply chain 3 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Resumen El sector agroalimentario es uno de los sectores más grandes e importantes de la economía industrial actual. Las pérdidas de producción y económicas en el mercado alimentario y a lo largo de la cadena de producción alimentaria son uno de los principales problemas a los que se enfrenta esta industria, principalmente debido a la falta de coordinación entre los diferentes actores involucrados en la cadena de producción (CP). En una CP, uno de los principales obstáculos para aumentar la eficiencia es la falta de un flujo integrado de información entre los diferentes proveedores y compradores del sector. Los socios de la CP del sector agroalimentario tienen diferentes experiencias y roles, lo que resulta en una asimetría lingüística que afecta el rendimiento de producción global debido a la pérdida de información en algunas partes de la cadena. La comunicación es crucial no solo de persona a persona, sino también de persona a máquina o de máquina a máquina. La Industria 4.0 y el Internet de las Cosas (IoT) ahora impregnan la mayoría del desarrollo de la CP, desde el campo hasta la mesa, encontrando también la introducción de ingenieros de TIC que introducen en la industria un lenguaje muy especializado para aplicar el desarrollo de herramientas de análisis. Estas herramientas (por ejemplo, modelos computacionales, evaluación del ciclo de vida) prometen un mejor rendimiento que ayuda a aumentar la productividad y reducir las pérdidas, introduciendo un lenguaje diferente. Por las razones mencionadas, es necesario definir un lenguaje común, y una de las soluciones estudiadas en los últimos años para mejorar la comunicación semántica se representa mediante las ontologías. Estas identifican y describen un vocabulario específico para un dominio, compuesto por términos anotados con una definición específica y relacionados entre sí. Las ontologías ayudan a vincular conceptos y procesos en la industria, contribuyendo a la comprensión y mejora de los sistemas complejos, incluyendo mercados y cadenas de suministro. El estudio de las ontologías se ha estudiado mucho en los últimos años y comienzan a tomar importancia como una forma de estandarizar la comunicación en la industria agroalimentaria. Este trabajo de grado comienza con una introducción a un análisis general de las cadenas de producción, junto con la relación con las cadenas de valor y sus diferentes objetivos. A continuación, se aborda un estudio de la estructura y funcionalidad de las ontologías basado en literatura, que nos permite definir qué son las ontologías, cómo y en qué ámbitos de la industria agroalimentaria se están utilizando, y cuáles son las herramientas más utilizadas para trabajar con ellas. Finalmente, se realiza un intento de creación de una nueva ontología para la cadena de producción del aceite de oliva, la cual se evalúa su utilidad y aplicaciones en el punto de discusión. An ontology - driven conceptual modeling of olive oil supply chain 4 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Table of contents LIST OF FIGURES ................................................................................................................. 5 LIST OF TABLES ................................................................................................................... 7 SYMBOLS AND ACRONYMS ................................................................................................. 8 ACKNOWLEDGEMENTS ....................................................................................................... 9 1. INTRODUCTION ............................................................................................................ 10 1.1 OBJECTIVES ............................................................................................................. 12 1.2 AGRIFOOD CHAIN CONCEPTS .................................................................................... 13 1.3 CONCEPTS RELATED WITH ONTOLOGIES ................................................................... 19 1.3.1 ONTOLOGIES IN COMPUTER SCIENCE ................................................................. 21 1.3.2. THE ROLE OF ONTOLOGIES IN SOLVING SOME OF THE AGRIFOOD CHAIN ISSUES .. 25 1.3.3. ONTOLOGIES FOR AGRIFOOD SUPPLY CHAINS .................................................... 30 2. METHODOLOGY AND APPLICATION ............................................................................... 34 2.1 OLIVE OIL CASE OF STUDY......................................................................................... 35 2.1.1 OLIVE OIL VALUE AND SUPPLY CHAIN ................................................................. 35 2.1.2 OLIVE OIL SC WEAKNESSES AND PROBLEMS ........................................................ 43 2.2. CONCEPTUAL FRAMEWORK ..................................................................................... 45 2.2.1 OOSC ONTOLOGY ENTITIES AND DEFINITIONS ..................................................... 47 2.2.2 THE GENERAL SUPPLY CHAIN ONTOLOGY SCONTO .............................................. 48 2.2.3 AN ONTOLOGY FOR OLIVE OIL SUPPLY CHAIN (OOSC ONTOLOGY) ........................ 57 3. RESULTS ....................................................................................................................... 58 3.1 OOSC PRODUCTION PROCESS DIAGRAM ................................................................... 58 3.2 OOSC ONTOLOGY VS UML MODELING ....................................................................... 60 3.3 EXAMPLE USE OF THE CONCEPTUAL MODEL ON PLASTIC ISSUE .................................. 64 4. DISCUSSION .................................................................................................................. 66 CONCLUSIONS .................................................................................................................. 68 BIBLIOGRAPHY ................................................................................................................. 70 An ontology - driven conceptual modeling of olive oil supply chain 5 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech List of figures Figure 1. Agri-food chains network diagram ……………………………….…………………………..………………11 Figure 2. Food Loss and Waste data of fruit and vegetables at different value chain stages during years 1972-2022 …………………………………………………………………………….………………………………………11 Figure 3. Objectives defined by steps in a diagram ………………………………….………………………………12 Figure 4. Simplified analytical perspective on food chains ………………………………………………………13 Figure 5. Simplified analytical perspective on supply chain and value chain ……………………………14 Figure 6. Monthly production and consumption of olive oil in Spain 2020 ………………………………16 Figure 7. Monthly production and consumption of olive oil in Spain 2021 ………………………………16 Figure 8. Diagram of the conceptualization of an ontology model ……………………………………………23 Figure 9. Basic components for ontologies conceptualization …………………………………………………23 Figure 10. UML model diagram example ……………………………………………………………………….………..23 Figure 11 ER-model diagram example ……………………………………………………………………….……………23 Figure 12. Been food product ontology represented by FOODON terms …………….………………… 24 Figure 13. Example of a blockchain ……………………………………………………………………….………………..27 Figure 14. The relationship between food waste, food safety and consumption at different stages of economic development ……………………………………………………………………….………………….27 Figure 15. Business views on challenges relating to complexity and uncertainty of processes atand behind-the-border ……………………………………………………………………….………………………………….29 Figure 16. Simplified System ecosystem of OBO Foundry ontologies used in FoodOn .……………31 Figure 17. “Extra virgin” olive oil long and short value chain ……………………………………………………37 Figure 18. Olive oil long value chain ……………………………………………………………………….……………….37 Figure 19. Material balance for the traditional (a) and for the two continuous three-phase (b) and two-phase (c) centrifugal systems ……………………………………………………………………………………39 Figure 20. Olive oil value chain ……………………………………………………………………….……………………….40 Figure 21. Activities carried out in the different olive oil value chain processes ……………………….42 Figure 22. Ontobee search for “olive” with different ontology entries an ID numbers ………………46 Figure 23. Ontobee “olive” class hierarchy ……………………………………………………………………………..47 Figure 24. SCOPRO structure concepts diagram ………………………………………………………………………49 Figure 25. Main classes of the processes inside a supply chain ……………………………………………….51 Figure 26. Specific concepts inside the process dimension of a supply chain …………………………..52 An ontology - driven conceptual modeling of olive oil supply chain 6 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Figure 27. Specialized classes of business processes in a supply chain ……………………………………..53 Figure 28. Temporal relationships of processes inside a supply chain in SCOPRO …………………….54 Figure 29. Resource classes and relations of the processes inside a supply chain …………………… 55 Figure 30. Olive oil Value Chain ……………………………………………………………………….………………………57 Figure 31-a. Decomposition of the olive oil production process of an olive oil company ………. 59 Figure 31-b. Decomposition of the olive oil logistic process of an olive oil company …………………60 Figure 32. OOSC ontology UML diagram ……………………………………………….………………………………..63 Figure 33. OOSC UML Diagram adding the cycle of plastics ……………………………………………….…….65 An ontology - driven conceptual modeling of olive oil supply chain 7 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech List of tables Table 1. Types of Olive Oil, based on the study of the “Agencia de Defensa de la Competencia de Andalucía”.……………………………...…………………………………………………………………………….………………35 Table 2. Types of Virgin Oils” ………...………………………………………….…………………………………..………36 Table 3. Olive oil supply chain processes descriptions …………………….……………………………………….40 Table 4. Principal entities of olive oil supply chain found in other ontologies ……..……..………….48 Table 5. Legend of symbols used in the entities of an olive oil supply chain process diagram ..63 Table 6. Arrow symbology relations explained for OOSC ontology …..………………….…………..…….58 An ontology - driven conceptual modeling of olive oil supply chain 8 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Symbols and acronyms SC = Supply Chain OOSC = Olive Oil Supply Chain VC = Value Chain OWL = Web Ontology Language RDF = Resource Description Framework GTTS = Global Track & Trace System BPM = Business process management EAGLET Ontology = Event AGent Location Equipment and Thing Ontology DSC Ontology = Dairy Supply Chain Ontology MESCO = MEat Supply Chain Ontology SCONTO = Supply Chain ONTOlogy) UML = Unified Modeling Language ER-Model or ERM = Entity relationship model An ontology - driven conceptual modeling of olive oil supply chain 15 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Food loss - All food and inedible parts of food removed from the food supply chain for recovery or disposal (including composting, under-ploughed/unharvested crops, anaerobic digestion, bioenergy generation, cogeneration, incineration, disposal to sewer, landfill, or ocean) (Östergren et al., 2014) by food suppliers in the chain, excluding retailers, food service providers and consumers. (SOFA, 2019) Food waste - Decrease in the quantity or quality of food resulting from decisions and actions by retailers, food service providers and consumers (SOFA, 2019). Market - The formal definition of the word market is: "A means by which the exchange of goods and services takes place by bringing buyers and sellers into contact either directly or through intermediaries or institutions." (Britannica, 2022). The market is the starting and ending point of all supply and value chains, it is the place where contact between producers and consumers takes place, and lies between the production chain. Market dynamical aspects - Markets do not always behave uniformly - fluctuations can be attributed to any disturbance, such as the occurrence of additional costs or retaliatory actions that cause a market participant to deviate from the usual market performance, also called externalities. An example of this undesirable behaviour would be the climate in fruit markets, which is constantly changing, leading to yield losses that throw market prices out of balance. In the agri-food sector, price fluctuations are a problem that prevents efficient planning of various process activities. The fruit market needs good planning that considers the fruit calendar, consumer demand and climatic conditions as much as possible (Figure 6.) (Figure 7.). This enables a reduction in the time divergence of the individual processes in the supply chain. An ontology - driven conceptual modeling of olive oil supply chain 16 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Figure 6. Monthly production and consumption of olive oil in Spain 2020 Source: Own elaboration with data based on MAPA Figure 7. Monthly production and consumption of olive oil in Spain 2021 Source: Own elaboration with data based on MAPA Externality - A consequence of an industrial or commercial activity that affects other parties without being reflected in market prices, such as the pollination of surrounding crops by bees kept for honey production. Oxford Languages, 2022. An ontology - driven conceptual modeling of olive oil supply chain 17 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Three types of externalities can be considered, affecting: consumers, producers and global common issues. ● Externalities that affect consumers: These are those that are not related to the production or consumption of the goods, but consumers are affected by external subjective added values of the product (self-belief values ). For example, sustainability or animal welfare: if consumers are concerned about the methods of meat production, they will only buy products that have been produced under certain animal welfare conditions. ● Externalities that affect producers: these occur when the production process of a good is altered by external forces that do not affect prices. Climate factors can be among these externalities, as, for example, water scarcity due to low rainfall can reduce grain yields and lead to a decrease in production and consequently an increase in grain prices. ● Global commons: this type of imperfection involves resources that are considered global commons (open access resources with no or undefined property rights), such as fossil fuels or edible fish in the sea. This external impact is usually the result of overuse of finite resources. For example, the decline of the tuna population results in a rising price in the fish market. Market control - A good/commodity market is subject to value/price fluctuations - agricultural markets may be driven by climate-related yield fluctuations or demand fluctuations. One strategy to reduce price volatility is consortia solutions, where a body sets policies and management strategies for the value and supply chain and achieves better communication to reduce the irregularities and delays between demand, production and consequently prices. At the territorial/regional level, control occurs at the political level, e.g., through government bodies or parallel structures, but these can lead to market imperfections, including: ● Monopoly - Market structure in which a single seller or firm faces many buyers in the market, making that seller or firm powerful enough to prevent competitors from entering the market. This results in the power to determine the final price of the product and limits consumer choice. An ontology - driven conceptual modeling of olive oil supply chain 18 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech An example related to food-agriculture products would be the monopoly of the world seeds market by the firms Monsantos, Corteva Agriscience and Syngenta that owned 65 percent of the international sales in 2017 (Sanchez et al. 2021). According to Philip Howard, when four firms control 40% of a market, it is no longer competitive. Other examples of monopoly in the agricultural market are agrochemicals and agricultural machinery equipment which in 2013 were led by only three companies comprising 51% (Syngenta, Bayer and BASF) and 49% (Deere, CNH and AGCO) respectively. (Grupo ETC, 2015) ● Monopsony - Market structure in which a buyer or a few buyers who have significant market power can determine the price of goods paid to a large number of sellers. It can also be called a monopoly, but from the buyers' point of view. In agriculture, there are many cases of monopoly because many small sellers face a single buyer or a small number of buyers. An example of this market behavior could be the case of the American company United Fruit Company (UFCo), founded in 1899 in the "banana republic" of Costa Rica. The company was the only employer in the region for the plantations and set the price for the producers (local farmers) without having competitors in the same region (MéndezChacón, 2022). Another example is sugarcane mills in Pakistan, where the price is set by the milling companies - they benefit from regulations that prevent farmers from taking the products to another district and from timing between crops to force farmers' land use decisions (Mustafa, 2023). Symmetric - One of the partners involved in the industry project has a different understanding of the data. Symmetric information in economics is one of the main requirements for the efficient functioning of competitive markets, and it helps to prevent sellers from extracting an excessive price from buyers if those buyers have an equal amount of information about a given product. (S. Bain, 2020) An ontology - driven conceptual modeling of olive oil supply chain 19 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Imperfect - Not all consequences that can create a market are known or have been considered. Individual buyers and sellers can influence prices and production, there is no full disclosure of information about products and prices, and there are high barriers to entry or exit in the market. (W. Kenton, 2020) Industry 4.0 - In recent years, technologies have become increasingly important in industry, with talk of Industry 4.0, a "period of significant transformations in product manufacturing using technology that harnesses the capabilities of computers and automation, augmenting them with intelligent and autonomous systems driven by data and machine learning." (European Innovation Council, 2018). This current shift to a cyber-physical system of the industry is another fact that could improve the efficiency of the food supply chain by helping to regulate all stages of production and monitor all processes in the chain. Considering that the industry is supported by technology and the processing of all data, a standardized language should be created between sectors and companies throughout the chain to avoid misunderstandings and misleading information. 1.3 Concepts related with Ontologies In this section an introduction of ontologies will be made so as to understand its part and know its purpose and functioning in industry and markets at the present time. Moreover, some specific examples will be presented to see where the actual investigations and uses of ontologies are leading in the agrifood chains sector. Definition - Ontology is a word with ancient origins that has been recently reused to denote tools for improving the language problem between different fields of knowledge. Originally developed in the life sciences, they are now receiving increasing attention in many fields, including the food supply chain. It has two different definitions regarding the sector it is used for: ● Lexico Semantic - The word "ontology" comes from Latin and is a translation from the Greek of the word "metaphysics", which means "above physics", in other words the universal discipline that describes the categories of reality. Or in the case of the An ontology - driven conceptual modeling of olive oil supply chain 20 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech separation of the word "onto-", which means "being", and "-logia", which means "logical discourse", as in the logical definition of a reality or a physical object. ● Computer science - In computer science, the definition of OWL (Web Ontology Language) states that an ontology is a "semantic language designed to represent extensive and complex knowledge about things, groups of things, and relationships between things. It uses a language based on computer logic so that the knowledge expressed can be used by computer programs" (W3C, 2013)(10). DEFINITION OF ONTOLOGY FOR THIE PRESENT WORK An ontology can be represented as a model that represents knowledge in terms of concepts within a domain while capturing the relationship between those concepts and linking them together to create an organized structured model that is easier to understand and manage. In this work scenario the ontology described and developed is represented by the computer science definition. History - Ontologies were first introduced in the 1970s as part of the new AI boom in computer science and described by Patrick Hayes as "logical formalization of common sense knowledge applied to robotics" whose goal was to transform human common sense knowledge into computer-readable knowledge. in 1990, Tim Berners-Lee created the Semantic Web, an Internet that captures content created using ontologies. in 1993, Tom Gruber added a new concept to ontological meaning by developing ontologies that explicitly used the framework of portable knowledge specification, which meant that knowledge could be used by other disciplines, not just in a specific language of one group. In 1994, Berners-Lee in turn created the World Wide Web Consortium (W3C), which defined the technical, academic aspects of the Web, as well as ontologies, to have a universal, standardized understanding to better manage data. In 1990, the Protégé ontology editing software was developed, which made it very easy to create ontologies, allowing for faster ontology development. A few years later, in 2004, the Web Ontology Language (OWL) was developed, which is now the main language for creating ontologies. An ontology - driven conceptual modeling of olive oil supply chain 21 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech 1.3.1 Ontologies in computer science Description 1 - Dynamic language that help improve communication between two or more actors and use semantics described as a content theory about the types of entities, their properties, and relationships in a particular domain of knowledge (Chandrasekaran et al. 1999) Each domain has its own ontology that contains potential terms describing the knowledge, processes, and actions of the subject in question. It is not the vocabulary as such that counts as an ontology, but the conceptualizations that the terms in the vocabulary are meant to capture, and the relationships between them. Ontology specification in knowledge systems has two dimensions: ● Expertise knowledge: provides knowledge about the objective facts in the domain of interest (objects, relationships, events, states, causal relationships, etc.). ● Problem-solving knowledge: provides knowledge about how to achieve various goals. Some of this knowledge might be in the form of a problem-solving method that specifies, in a domain-independent way, how to achieve a class of goals. Description 2 - Shared and common understanding of a specific knowledge domain that can be communicated between people and computer-based systems, as well as help computer understanding and processing information more efficiently. Providing vocabulary to describe data with semantics that computers can understand. There is a range of different languages designed to perform ontology for semantic web, RDFS and OWL are the most basic languages which are used commonly. (Hoang et al. 2020) An ontology - driven conceptual modeling of olive oil supply chain 22 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech OWL - Web Ontology Language is a language for defining ontologies on the Web. Describes a domain in terms of classes, properties and individuals and may include rich descriptions of the characteristics of those objects (Bechhofer et al. 2009). It is intended to facilitate interpretability among web content using vocabulary and formatting that allows automatic machine processing. (Rouse et al. 2016) RDF - Resource Description Framework, is a model for encoding semantic relationships between items of data so that these relationships can be interpreted computationally. (Oxford Lenguages) RDFS - Resource Description Framework Schema is a vocabulary, in RDF, that explains how nodes of graph relate. (W3C,2005) can define the vocabulary, specify which properties apply to which kinds of objects and what values they can take, and describe the relationships between objects. (Martinez et al. 2016) Hierarchy and representation An ontology is a conceptualization of an abstract model representing a domain by identifying the relevant concepts and how they are related with each other (Figure 8). The main components to represent ontologies models are (see Figure 9): ● Classes (or entities): sets of things that represent concepts. They can be concrete objects of the real world or abstract concepts. The description of these classes is described by attributes in a formal specification. ● Relations (or roles): special attributes, whose values are objects of (other) classes. ● Constraints (or rules): determine allowed values. ● Axioms: describe knowledge that cannot be expressed simply with the help of other existing components. ● Individuals: individual entities, represent concrete and abstract objects An ontology - driven conceptual modeling of olive oil supply chain 23 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Figur 8. Diagram of the conceptualization of an ontology model Source: Own elaboration based on lecture (Sack et al. 2020) Figure 9. Basic components for ontologies conceptualization. Source: Own elaboration based on lecture (Sack et al. 2020) Ontologies can be modeled via database or software modeling technologies, the most used are: UML (Figure 10) and ER-Model (Figure 11). Figure 10. UML model diagram example Figure 11 ER-model diagram example An ontology - driven conceptual modeling of olive oil supply chain 24 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech UML - Unified Modeling Language (UML) is a graphical language for visualizing, specifying, constructing, and documenting the artifacts of a software-intensive system (Gogolla et al.2009). ER-model - Entity relationship model is a conceptual model that represents the information structure of a problem domain in terms of entities and relationships (Gogolla et al.2009). Simple ontology examples As mentioned earlier, ontologies are divided into domains of knowledge. Each ontology may deal with the same concept, but from different points of view. In Figure 12 we have an example of food domain ontology. Figure 12. Been food product ontology represented by FOODON terms Source: CGIAR FoodOn is a comprehensive ontology with controlled vocabulary, built to interoperate with the international OBO Library collaboration and to represent “food role” entities allowing a farm to An ontology - driven conceptual modeling of olive oil supply chain 31 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech This ontology is based on the OBO Foundry source, a set of interoperable, life science-oriented ontologies with support from FAIR data sourced from academic research, government, and the commercial sector (Figure 16). Reused knowledge includes environmental terms from ENVO, agricultural terms from AGRO, plant and animal anatomy from UBERON and PO, organisms from NCBITaxon, relations from RO, and nutrient components from CDNO. Figure 16. Simplified System ecosystem of OBO Foundry ontologies used in FoodOn Source: Foodon 2021 Motivation: the main goal of FoodOn is to build a data sharing network to support the lack of data and data sharing between agencies and companies by combining all existing data connected to the Internet. The domain of this ontology is human food as it moves from its origin in nature or on the farm through the processing and distribution chains to the consumer. It follows the stages from farm to consumer. (Dooley et al. 2018) Benefits: Help solve data harmonization problems in food safety, quality, production and distribution, and consumer health and convenience. Improve system interoperability, quality control, and software-driven intelligence. Enabling IoT-connected transaction systems for food production, processing, and distribution to reduce data exchange costs. An ontology - driven conceptual modeling of olive oil supply chain 32 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech EAGLET Event AGent Location Equipment and Thing Ontology: is an ontology for individually identified goods in the supply chain to support decision making and communication while meeting two specific requirements, economic and collaborative (Geerts et al. 2014) Motivation: to use ontologies as a critical component of a supply chain of things to facilitate visibility and interoperability of things along the supply chain. Define an ontology that leverages the availability of a single thing's (object's) identifying information in the context of a standard set of economic phenomena that support multiple views in a range of data architectures. Define structuring principles that provide guidance for supply chain system design. Benefits: Increased visibility and interoperability of individual objects that facilitate management and collaborative decision making related to supply chain activities. DSC Dairy Supply Chain Ontology: ontology to improve DSC management efficiency by facilitating interoperability within DSC, mapped to the structures of OWL (Pizzuti et al. 2017) Motivation: enabling an increase in production productivity, improving planning and forecasting, and improving the competitiveness of dairy products. The DSC ontology aims to improve the efficiency of DSC management by enabling interoperability within DSC by helping to provide the basis for a transparent DSC model to improve supply chain management in terms of product-oriented safety (quality), visibility, optimization, and sustainability. Benefits: Creation of an ontology to define DSC domain knowledge to support data interoperability within the dairy ecosystem and improve DSC management efficiency. In addition, the proposed ontology would facilitate the application of IoT solutions and improve the applicability of sustainability-oriented DSC management solutions that comply with the recommendations of the European Dairy Association (EDA) and the International Dairy Federation (IDF). MESCO MEat Supply Chain Ontology: ontology developed to support meat traceability management and enable interoperability between different systems throughout the supply chain, using the standard language OWL (Jachimczyk et al., 2021) Motivation: to ensure food quality and safety and improve the lack of common standards for information encoding and management and the inability to link food An ontology - driven conceptual modeling of olive oil supply chain 33 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech products. To develop an ontology to support the management of meat traceability along the entire supply chain, from farm to final consumer. Benefits: Dealing with different traceability scenarios and problems. Enabling interoperability between different systems and integration of heterogeneous databases used by different actors in the supply chain. Facilitate the acquisition of essential data that is fundamental in the event of a foodborne illness outbreak, easily addressing the main issues of food safety officers. SCONTO SCONTO (Supply Chain Ontology) is an ontology that describes a general vocabulary for the supply chain. Sconto mainly aims at describing processes, performance evaluation and benchmarking and gathers most of the SCOR model developed by the Supply Chain Council (Harmon, 2019), whose value in terms of ontologies has been already analysed by Sakka et al (2010) and Rosing et al. (2015). SCONTO will be described in the next chapter and will serve as the basis for specific applicable ontologies. Motivation: to support the integration of material, information and monetary flows in the supply chain of companies in order to achieve a seamless semantic integration of Industry 4.0 and its digitalization. To achieve this, SCONTO is creating a main module called SCOPRO , which is an ontology with the goal of specifying the business processes associated with the supply chain based on sharing a precise meaning of the information exchanged between the many actors involved. Benefits: Description of the structure of the supply chain (SCONTO). In the case of the SCOPRO module, supply chains consisting of one or more companies can be accurately described by representing the structure and the organizations involved, specifying the organizational processes and relating the sub-processes to each other, describing the impact of the activities and multifunctions of the resources, and representing the material movements along the chain. With this information, a common vocabulary is achieved that is valid for different industries and also different scales to achieve the semantic integration of supply chains. An ontology - driven conceptual modeling of olive oil supply chain 34 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech 2. Methodology and application As mentioned in the objectives (point 1.1), this study wants to optimize an agrifood chain, in this case olive oil chain with the use of ontologies. The methodology used to create the final olive oil SC ontology starts with a proper gathering of detailed information of olive oil SC processes, actors, weaknesses and problems. The information has been collected with the following methodologies: ➔ Review of bibliography ➔ Interviews to different sector experts ➔ Attendance to workshops about identifying some of the sectors problems ➔ Collection of data from public government statistics Interviews with sector experts The carried-out interviews have been performed with 2 teams of experts in worldwide (17 of February of 2023) and Catalan olive oil SC (25 of April of 2023) from Italy and Catalonia. The content of the interview included questions to understand: • The functioning of the sector in each step (agronomic, production, distribution or consumption). • Detection of problems that may be related to production and/or economic losses. • The functioning of distribution and export logistics • The functioning of market operation • How is consumers perception of the product. Workshop about plastic in the agrifood value chain Along with the thesis process research of problems inside olive oil SC, the circular management of plastics held by the new policy of plastics in Catalonia has been one of them. In order to understand and approach this topic, I was able to attend a workshop held by CREDA to encourage the transfer of knowledge and innovations that help circular management of plastic waste in the agri-food sector in the Penedès. Where different actors of the agrifood SC attended the session to evaluate which stage of the chain would be more affected by the reduction of plastics and propose solutions in order to do so. An ontology - driven conceptual modeling of olive oil supply chain 35 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech 2.1 Olive oil case of study In this section an exhaustive analysis of the olive oil production chain will be performed to understand the actors and relations of their processes in order to achieve a well-established ontology. Furthermore, a research and analysis of chain weaknesses and problems will also be performed to try to represent them in the ontology to see the applicability of problem solving of the model. 2.1.1 Olive oil value and supply chain In the past, the production chain of olive oil was considered simple compared to other agricultural products, since it consisted of collecting olives, storing them and processing them into oil. Thanks to new technologies, a better system of oil extraction and consequently a better quality of the product has been achieved, making the olive oil production chain more complex. Olive oil is divided into different product categories that are distinguished by certain characteristics, the most important of which is the maximum free acidity. The categories are first divided into three main types of oils: "Virgin Olive Oils", which include Virgin Olive Oil, Extra Virgin Olive Oil and Lampante Oil; "Refined Olive Oils"; and "Olive Oil" Each category follows the basic process but with small changes that repercuss on the final quality and properties of the product. Table 1. Types of Olive Oil, based on the study of the “Agencia de Defensa de la Competencia de Andalucía”. Virgin Olive Oils Oils obtained from the fruit of the olive tree solely by mechanical processes or other physical processes under conditions which do not cause alteration of the oil and which have not undergone any treatment other than washing, settling, centrifugation or filtration. Refined Olive Oil Similar to an olive oil obtained by refining virgin olive oils, it has a free acidity of no more than 0.3 g per 100 g of oleic acid. Olive Oil Oil containing only refined olive oil and virgin olive oil. Made from a mixture of refined and virgin oil, without lampante. The content of free fatty acids is not higher than 1 g per 100 g of oleic acid. An ontology - driven conceptual modeling of olive oil supply chain 36 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Table 2. Types of Virgin Oils Virgin Oils Virgin Olive Oil Presents a maximum free acidity of 0,8 g per 100 g OA* Extra Virgin Olive Oil Presents a maximum free acidity of 0,8 g per 100 g OA* Lampante Olive Oil Presents a free acidity not higher than 2 g per 100 g OA* *OA: Oleic acid 2.1.1.1 Olive oil value chain For the analysis of the olive oil chain, it is necessary to understand that there are two independent value chains, "extra virgin olive oil" and "olive oil", which have an impact on costs and prices due to their different production processes. In the chain, three major processes or phases are distinguished: production (primary activity), processing (or industrial phase) and distribution or marketing. Depending on the degree of extraction or complexity of processing required for each type of oil, a distinction can be made between two channels, the short and the long (Figure 17) (Figure 18). The long channel is used by the major distributors and marketers who distribute "extra virgin" and other olive oils. The short channel, on the other hand, is used exclusively for "extra virgin" olive oil", which is marketed with its own quality label. Channel - Logistics channel refers to a network that includes all supply chain participants involved in functions such as transportation, receiving, handling, storage, and information exchange. MONASH University, 2023 An ontology - driven conceptual modeling of olive oil supply chain 37 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Note: “Extra virgin” olive oil does not undergo a refining process or the mixing with other olive oils of different acidities, so the refineries do not exist in this channel. Figure 17. “Extra virgin” olive oil long and short value chain Source: Own elaboration based on the study of the “Agencia de Defensa de la Competencia de Andalucía”. Figure 18. Olive oil long value chain Source: Own elaboration based on the study of the “Agencia de Defensa de la Competencia de Andalucía”. An ontology - driven conceptual modeling of olive oil supply chain 38 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Description of the activities done in each phase: ● Production: olives are harvested from the field and transported to the oil mill, where they are processed into oil by olive growers, individuals or farmers united in a cooperative, who are responsible for growing and harvesting the olive trees. ● Processing: industrial process with the oil mill as the main actor, playing the role of milling and storage. In Spain there are two legal forms of oil mills: Cooperatives that grind only the olives of their partners. Industrial oil mills: Industries or private companies that grind the olives of the producers through a contract. Oil mills sell the oil mainly through two channels, farmers' own consumption or local markets (virgin and extra virgin) and refineries (lampants), packaging plants (virgin and extra virgin) and to operators. Olive oil must also go through the process of refining and packaging plant. To obtain olive oil, lampante olive oil must be refined and later blended with virgin oils before it can be packaged and marketed. Most refineries belong to industries that also perform the packaging process. There are three types of packers: "packers integrated with refineries" (sell all types of olive oil), "packers belonging to oil mills" (pack only extra virgin olive oil), and "independent packers" (pack all types of oil). In some cases, a broker is involved in the chain, handling the sale of bulk olive oil, working as an intermediary between oil mills and refineries or packers, and receiving a commission from the total revenue. ● Distribution or commercialization: this is the last stage of the oil chain before the oil reaches the consumer. The bottled oil reaches the purchasing centers or distribution platforms, where it is managed and transported from the refinery or bottling plant to the various points of sale (logistics for receiving and delivering the product). In the case of extra virgin olive oil (short distribution channel), the bottling plant is also the distributor of the product. Olive Oil chain residues The chemical composition of olive is about 50% moisture, 22% oil, 19% sugar and 9% other organic compounds (cellulose, proteins, ash). Thus, during the production of olive oil, two main An ontology - driven conceptual modeling of olive oil supply chain 39 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech residues are produced: Pomace and plant water, which are produced after crushing, malaxation and finally decantation of the olives in the oil mill. ● Pomace: solid residue obtained during olive oil production by pressing or centrifugation. ● Plant water: water containing solid organic compounds dissolved (1%), derived from the moisture of the olives and other added water used in olive oil production. Figure 19. Material balance for the traditional (a) and for the two continuous three-phase (b) and two-phase (c) centrifugal systems Source: G. di Giacomo et al 2022 The treatment of the residues is compulsory. The plant water has a high pollution risk due to its high levels of polyphenols which are toxic, the available treatments to reduce the content of contamination of this water are by anaerobic digestion or aerobic treatment (Ibañez et al. 2020). For the subproduct pomace, it is brought to the secondary extraction factories where the pomace oil is extracted. The final residue of the process of pomace oil is used for energy production or composting material. 2.1.1.2 Olive oil supply chain Olive oil supply chain (Figure 20) generally consist of three stages: 1. Olive production 2. Olive oil processing (mills, packing plants, refineries) 3. Distribution (small sales to the consumer, restaurants, supermarkets) Between these stages there is transportation, which may be of great or little importance depending on the proximity of the places where the stages are placed. In Table 3 we describe all the processes that take part on these three stages. An ontology - driven conceptual modeling of olive oil supply chain 40 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Figure 20. Olive oil value chain. Source: MARM, 2011 Table 3. Olive oil supply chain processes descriptions Source: MARM 2011 PRODUCTION Olive Grower Farmers who can work alone or in business associations (cooperatives or agricultural partnerships, known by the Spanish acronym SAT). They grow the olives, harvest them and transport them to the oil mill where the olives are pressed and the oil is extracted. Types: Traditional or intensive, organic or integrated cultivation PROCESSING Olive Oil Mill Machine or factory in which olive fruit is crushed or pressed to obtain olive oil. Types: Cooperatives, agricultural partners or privately owned mills. Sales channels: for farmers' own consumption or in bulk to refineries, packers or wholesalers. Refineries Olive oil factory where heat, chemicals, and other processes are used to change the flavor, color, and other properties of the oil. This is done only with lower quality olive oils; higher quality olive oils (extra virgin) are not refined before consumer use. Packing Plants Companies that package the product at different stages of the olive oil value chain. The types depend on the stage of packaging: Packers integrated with refineries, packers belonging to large mills or second level cooperatives, and independent packers. An ontology - driven conceptual modeling of olive oil supply chain 47 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Note: “entity” is the highest super class Figure 23. Ontobee “olive” class hierarchy, Source: Ontobee (2023) YEd program yEd is a powerful desktop software written in Java that can be used to generate high-quality diagrams quickly and effectively. Create diagrams manually or import your external data for analysis with automatic layout algorithms that arrange large data sets. It contains UML diagram features amongst others that enables to create and design models. 2.2.1 OOSC Ontology entities and definitions The first step of ontology creation is based on the collection of vocabulary within the chosen domain. To obtain a universal base vocabulary, a top-level ontology must be used, in this case OWL, to connect with other ontologies. To search these already defined entities related to Olive Oil SC we used Ontobee and found the principal actors shown in Table 4. This entities ID and descriptions will be reused in the new olive oil supply chain ontology, whereas the entities without equals in other ontologies they will need to be defined. Find in Annex A the rest of the ontology entities found in Ontobee website. An ontology - driven conceptual modeling of olive oil supply chain 48 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Table 4. Principal entities of olive oil supply chain found in other ontologies. Source: Ontobee website Entity Ontobee Ontology ID Description Orchard ENVO ENVO_0000 0115 An intentional planting of trees or shrubs maintained for food, typically fruit, production. Olive oil FOODON FOODON_0 3301826 Undefined Farmer AGRO AGRO_0000 0372 A person engaged in agriculture, raising living organisms for food or raw materials. Mill — — — Packing plant — — — Distributor NCIT NCIT_C4527 4 Indicates the person or authoritative body who has distributed something. Supermarket NCIT NCIT_C1806 6 A large self-serving retail or wholesale market that sells food and household goods. Consumer — — — 2.2.2 The general Supply Chain Ontology SCONTO As mentioned in 1.4.3, the SCOR standard by the SC Council (Harmon, 2019), has been recently transformed in the ontology (Sakka et al,2010; Rosing et al.,2015). Here the Vegetti’s (2021) approach has been used, SCONTO, developed accounting for the following points: 1. Requirements specification: identifying the scope and purpose of the ontology. 2. Conceptualization: organizing the informal ontology view into a semi-formal specification (UML diagrams and OCL specifications). 3. Implementation: codification of the ontology in a formal language. 4. Evaluation: technical assessment, quality and usefulness of the new ontology. UML Diagram - Unified Modeling Language diagrams is a conceptual model which is often used for designing the logic-model of information systems. Hoang et al. 2020 An ontology - driven conceptual modeling of olive oil supply chain 49 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech OCL specifications - Object Constraint Language is language that gives the ability to precisely specify the properties of your system. It is complementary to the diagram and was designed to provide more details into the design of systems. Rugaber et al. 2015 SCONTO use textual expressions for entity definitions and their relationships and constraints, which can be visualised by UML class diagrams. In this section only entities are described, while relations and constraints are reported in Annex C. The whole picture of main entities is represented in figure 24, which puts in evidence three different dimensions: structure, processes and resources. Figure 24. SCOPRO structure concepts diagram Source: Vegetti et al. 2021 Structure Dimension - The structural dimension includes the Supply Chain itself and the market. Definition 1. The Supply Chain class represents a network of Organizational Units (OUs) that transforms or adds value to materials, ranging from raw materials sourcing,to the final product distribution in specific markets. The relations in which this class participates and its constraints are listed in Table 5 of Annex C. An ontology - driven conceptual modeling of olive oil supply chain 50 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Definition 2. A Market is a set of actual and potential buyers that are grouped together because they share certain distinctive characteristics. More general entities used to describe and detail structure include the above mentioned Organizational Unit: Definition 3: An Organizational Unit (OU) represents the enterprises or enterprise components that participate in the SC operation. Its relationships are described in Table1 of Annex-1. The OU include the subclass Functional area Definition 4. The Functional Area class captures the function-based work division that often occurs within organizations. This concept is further specialized into Purchasing, Production. Logistics, Marketing, Sales, R&D and Finance. Table 3 of Annex-1 presents the relationships in which this concept participates and its restrictions. In some context the entity Functional SubArea is introduced which groups the activities that are part of a Functional Area and that deals with a type of common tasks within the functional area of which they are part. Table 4 - Annex 1 presents the relationships in which the Functional Sub Area concept participates and its restrictions. Introducing Roles - An OU is an agent interacting with others - SC and Market are entities emerging from such interaction when the interaction generate a self-sustaining dynamical system. Each agent is represented by an actor who is given a role. Definition 5. An OU Role represents the functionality assumed by an organizational unit when participating in a given supply chain. An ontology - driven conceptual modeling of olive oil supply chain 51 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech The OU Role includes the following subclasses: ● Primary Producer Role: it is assumed by an OU that carries out activities related to the primary production, such as agriculture, livestock, fishing, mining, fossil fuels extraction, etc. ● Secondary Producer Role: role that is played by an OU when performing activities where raw materials or intermediate products are transformed into higher-value products (e.g. oil). ● Customer Role: a role that is performed by an OU that carries out activities related to the purchase of goods for use, consumption or exploitation. ● Service Provider Role: a role that is played by an OU that only performs activities related to transportation, marketing, storage, etc., that add value to a product without transforming it in a physical and/or chemical way. Relationships and constraints of OU Role are reported in Table 2 of Annex-C. Process dimension - Activities needed for a detailed description of processes, the temporal relations between them, the resource participation and the occurrence involved. The main concepts belonging to this dimension are: Process, Process Occurrence, Business Process, Process Element, Task, Temporal Relationship and other utilization concepts. Related classes are shown in Figure 25. Figure 25. Class diagram about Process Dimension Source: Vegetti et al. 2021 An ontology - driven conceptual modeling of olive oil supply chain 52 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Definition 6. A Process represents an activity chain that is carried out in a SC to achieve certain results. A process execution implies the creation, modification, use or movement of several resources, which may be physical or conceptual ones. In order to properly describe a SC, the business processes belonging to it should be modeled at different levels of abstraction. SCOPRO represents the process decomposition into more specific activities using the isSubProcessOfassociation, which is shown in Table6-Annex-C. Definition 7. The Process Occurrence concept represents a particular execution of a process in a given period. Based on the SCOR model (Supply Chain Council, 2012), SCOPRO considers three different types of processes that are needed to model and analyse a SC. The three concepts/entities that specialize the Process class are introduced in Figure 26, and are labelled as Business Process, Process Element (PE) and Task Figure 26. Process subclasses Source: Vegetti et al. 2021 Definition 8. A Business Process (BP) is a type of process that is composed of value-added activities. It is designed for achieving a result having a significant impact on clients and in the efficient management of SC flows. Each Business Process is composed of Process Elements. An ontology - driven conceptual modeling of olive oil supply chain 53 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Definition 9. A Process Element (PE) is an activity or a logical structure of activities that is part of a Business Process. The detail level that is provided by the Process Element class allows the decomposition of a Business Process into specific operations but, at the same time, is general enough to describe activities that are valid for different types of supply chains. Definition 10. A Task is an activity or a logical structure of activities that is part of a Process Element in a certain SC. Table 7 Annex-C presents the relations and constraints related to these concepts. SCONTO extends the Business Process and Process Element concepts using the vocabulary of the SCOR using the main business processes: Sources, Make Deliver, Source Return (to provider) and Deliver Return (from clients). The model also includes the planning of the operational activities related to material transformations and movements (Plan), as can be seen in Figure 27. Figure 27. Business Process subclasses Source: Vegetti et al. 2021 An ontology - driven conceptual modeling of olive oil supply chain 54 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Definition 11. The Utilization concept represents the way in which a process affects a resource that participates in a given component (subprocess, activity) of such process: Creation, Elimination, Modification, Use, and Material Transfer. Table 8-Annex C presents the relationships associated with the Utilization class. Timing - Together with the relation and constraints between classes of the same or different dimensions, a particular set of entities has been introduced which takes into account for time relations. Definition 12. Temporal Relationship represents different constraints related to the partial order between the executions of two processes in a SC. Figure 28 also show the different partial order relations associated with the execution of two processes., which are also part of a hierarchical structure. SCOPRO considers two types of temporal relation: atomic and composite ones. The former includes “Before than”, “Meets”,”Overlaps”,”Equals” and “During”. Figure 28. Temporal relationships of processes inside a supply chain in SCOPRO. Source: Vegetti, 2021 Process is also related to the usage of resource, through the concept of utilization. An ontology - driven conceptual modeling of olive oil supply chain 55 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Definition 13. The Utilization represents the way a process affects a resource that participates into a given activity: Creation, Elimination, Modification, Use, and Material Transfer (see Figure 24). Table 8 - Annex C -presents the relationships associated with the Utilization class. Resource dimension - Concepts specifying the resources and the role they play when participating in a process. The main concepts belonging to this dimension are: Resource and Resource Role (see Figure 29). Figure 29. Resource classes and relations of the processes inside a supply chain. Source: Vegetti, 2021 Definition 14. A Resource can be any type of physical or conceptual medium that participates in a process via one of its roles. This class represents resources that flow through the network (e.g., commercialized goods) and those that remain static (e.g., industrial plants). Supply Chain resources include buildings or facilities, material handling equipment, different types of products, information and financial resources, among others. Therefore, SCOPRO specializes the Resources lass into the Material Resource, Information Resource, Financial Resource, Human Resource and Facility classes, which are further specialized. Definition 15. A Resource Role reflects the participation type that a resource has in a given process. An ontology - driven conceptual modeling of olive oil supply chain 56 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech The same resource can play different roles in distinct processes. For example, a drill can be the final product of a production process and can be a tool in another one. Table 9 -Appendix C presents the relationships associated with the Resource Role class, as well as its constraints. Limits of SCOR & SCONTO approaches In the bottom side of diagram of Figure 24, it can be noted some former limits of the approach which delegates to the concept of creation most of activities that in an agri-food chain are hard to be understood, as they lack the concept of transformation. Also, the concept of elimination can be criticized from the viewpoint of waste management, LCA analysis and circular economy. In the diagram of Figure 27 we find the same conceptual limitations found in the process subontology. Here source and make reflects the concept of creation, to be better meant as a product development. In the class diagram of Figure 28 it appears how SCOR (and SCONTO) put a major attention to temporal synchronization of Processes. In fact, in Vegetti the case study is about the delay occurring from market demand to product availability. From the diagram of Figure 29 it can observed a limited detail spent on then description of resource, which doesn’t include products, goods, services, etc., Practical use of SCONTO - a dictionary of symbols While the SCOR model can be substantially rendered by UML class diagrams, to use the entities to describe a system, it can be useful to coin a new symbol dictionary. Here is the legend of symbols (slightly different from the one used by Vegetti) representing a selection of the entity described above (Table 5), which may be used to facilitate the description of a business process. Table 5. Legend of symbols used in the entities of an olive oil supply chain process diagram. An ontology - driven conceptual modeling of olive oil supply chain 63 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Figure 33. OOSC ontology UML diagram Source: Own elaboration based on information recorded in point 2.1 and 3.1 Figure 32. OOSC ontology UML diagram Source: Own elab2ration based on information recorded in point 2.1 and 3.1 An ontology - driven conceptual modeling of olive oil supply chain 64 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech 3.3 Example use of the Conceptual Model on Plastic Issue Recently the OOSC is experiencing a new issue, represented by the use of plastic bottles in oil. In fact, a new Spanish regulation, pushed from EU recommendation, is making it mandatory for producers and bottlers to zero the use of plastic in a few years - fines are expected to be applied to every country not attaining such directions. In the model depicted in Figure 33 the estimators of the use of plastic resources that take part in the OOSC are circled in red. The diagram describes the flow of the material resources in the chain (oil, plastic) also involving the final packing process. It is possible to observe the close link that exists between certain supply chain business processes involving different organizations. The modelling of the process interactions is an essential step to achieve semantic interoperability between SC information, the design of information systems that support SC collaborative management, as well as the generation of information systems supporting material traceability and allowing a comprehensive visibility of the information. In Figure 33 we can also see that there are two principal actors involved in the use plastic. The first one is the farmer” that by means of olive orchard’s agronomic inputs, indirectly (not found in end product) and in a reduced amount, use plastic containers of phytosanitary treatments or fertilizers. A more direct (found in end product) use of plastic is also used in of the watering pipes used in irrigated orchard. The second main actor, that use directly plastic, is the “packing plant” which in Spain uses mostly plastic bottles as containers for the olive oil. The cycle of plastic (in blue) of the packaging containers could be calculated by the model from the first initial stage to the last one end life of the plastic. It also could be compared with other alternatives (in green) to evaluate which is best for the business. It has to be emphasized that this representation is only a simplified hypothetical case of the olive oil plastic cycles inside the model, as it is a complex task that requires a lot of time to be finished as it needs a lot of reasoning and information. This model can also be adapted to the requirements of the analysis that needs to be done as more classes and attributes can be added depending on the information that is required. An ontology - driven conceptual modeling of olive oil supply chain 65 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Figure 34. OOSC UML Diagram adding the cycle of plastics Source: Own elaboration based on information recorded in point 2. Figure 33. OOSC UML Diagram adding the cycle of plastics Source: Own elaboration based on information recorded in point 2. An ontology - driven conceptual modeling of olive oil supply chain 66 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech 4. Discussion In the present study an introduction to the functioning and uses of ontologies has been made to develop an olive oil supply chain ontology conceptual model. The final conceptual model is represented by diagrams shown in Figure 31a, 31b, 32, 32. They include a description of the involved processes including milling, packaging and logistics, characterising companies with a different size. Figure 32 represents a general olive oil conceptual model based on a UML diagram that frames the big picture of all the stages and its actors with their main attributes and relationships adding different business contexts and equipment involved; also, an interpretation of the plastic cycle case recycling the previous UML olive oil chain diagram (Figure 33). The process diagram based on the classes defined from the previous studies and the dimensions defined by Vegetti et al. 2021, disaggregates the business context of an olive oil company above all the process of milling, packing and logistics in specific actions allowing to understand the detail interactions of each actor emphasizing the relevance of some resources used in the processes. The general OOSC ontology UML diagram gives a general view of the interactions of the actors that takes part in the context of the olive oil production helping to see the relationships once defined the classes and its attributes and detect the key points of the chain, being the olive orchard, mill, packing plant and customer. If transformed into software, the “methods” sections represented in the general ontological-model diagram will be the representation of the data calculation function that will be applied into computing to make the dynamical analysis. Today several problems are appearing that claims for a clear understanding of systems our society rei upon, as climate change and related extreme events, bringing the eve of new age where we should be prepared to find solutions to gain resilience by rapid innovation. The case of plastic a real issue that need for a quick solution. Last diagram dissects the cycle of plastic showing the two main actors that could suffer more from the new regulation, the farmer because of the use of plastic in agricultural management, and the Olive Oil producer because of the extensive use of plastic bottles. An ontology - driven conceptual modeling of olive oil supply chain 67 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech The OOSC ontology conceptual model is an innovative proposal, as only one attempt to use this approach for Olive Oil has been found to date (Guido et al. 2020), which analyses the olive oil SC in a simplified way and specifically on traceability aspects. On the other hand, the one derived in this study is a general base ontology aimed at covering the full chain. Other ontologies have been developed similarly to the OOSC ontology but in other agrifood domains as for example dairy (Pizzuti et al. 2017) and meat (Jachimczyk et al. 2021) Having created an OOSC ontology implies that if further steps of creating a computational model are accomplished, the supply chain could be able to embrace different improvements regarding the use of dynamical analysis, in aspects such as optimization of business processes management, predictions, traceability, reducing trading costs, efficiency of the resources, and so on. It also may be helpful to derive the architecture of other agrifood supply chains and the network including all the agrifood industry for a wider scale of optimization objectives. Ontologies are an innovative field of research applied to industry and have a little bibliography. Also, they are not many fully developed functional ontologies used in data analysis understanding, creating and explaining this study process has been difficult technically. Furthermore, ontology's purpose is difficult to define as its vocabulary and relationships can be constantly changing depending on the industry evolution. In addition, the creation of an ontological model is a complex task, as it is very important to understand all the processes that wants to be transformed into a model, including relationships and attributes of each actor and how they relate to different contexts. An ontology - driven conceptual modeling of olive oil supply chain 68 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Conclusions From the study of ontologies and the olive oil supply chain the first olive oil supply chain ontology has been proposed to contribute in formalization of the agrifood supply chain domain. OOSC ontology provides the bases for describing the three main stages of the chain, being olive oil production, processing, and distribution. Furthermore, a detailed decomposition of the milling and packing process of an olive oil company has been performed to provide understandable knowledge of the strong connection between these two industrial stages in olive oil production. This ontology general utility is to enable transparency in olive oil SC and interoperability between heterogenous data sources all along the chain to upgrade management efficiency and resources traceability. Complementing the study, an analysis of the cycle of plastic in the OOSC represented in the ontology model. This provides us an idea of the use of this ontology. to trace, detect the parts of the chain that are more affected by this resource. From this information we can make an overview of the repercussions of using this material. OOSC ontology provides the following capabilities: ● Introduce an understandable domain knowledge for olive oil SC. ● Represent semantic relationship between the actors of the three main stages of the olive oil supply chain, olive production, processing, and distribution. ● Emphasise the olive oil internal process knowledge of the stages of milling and packing. ● Base structure applicable to other agrifood supply chains. ● Capability of introducing olive oil process semantics to machine learning database. ● Database mapping for easier information search capabilities. ● Base diagram model to create dynamic analysis inside the olive oil supply chain to predict or help to make decisions in business. ● Explicit representation of plastic movements between olive oil supply chain partners. ● Detect resource material cycles such as the one of plastics to find and analyse strategic points that allow the optimization of the cycle by reducing or finding alternative solutions. An ontology - driven conceptual modeling of olive oil supply chain 69 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Ontologies can be a good base for helping to optimize the agrifood industry, as they can provide a tool to achieve a fully understanding of stages processes and reach common communication vocabulary amongst the diverse industry actors to avoid loss of information. This study is the start point in the development of a dynamic analytical model for the olive oil supply chain. OOSC ontology UML Diagram being the basis for the final computational model. This enables the creation of dynamic analytical data and statistics of this agrifood chain, meaning analysis with the advantage of having instant up to date data and historical data altogether. Which opens the door to detect new variables to understand and help to detect affectations on the chain, making the analysis algorithm learn and deduct from new data introduced improving the predictions. LIMITATIONS ODF THE CREATION OF THE CONCEPTUAL MODEL In the last years the ontological approach has grown and the toolset to work with them has become very complex. While from the one hand they may be used to develop a more reliable simulation models, becoming of strong interest of academic world and applied research, the complexity itself prevented to complete the path work to the reach of a computational model. Finally, we emphasise that investigating a topic that is still in development by experts and industry can be difficult to be achieved, as defining the results and evaluating them can be confusing. Apart from this small drawback, ontologies and its applicability in improving the optimization of the agrifood industry is a very interesting line of study that in a short time will be up to use in data analysis and AI software to facilitate the management in the agricultural and food sectors preventing food and economic losses. RECOMMENDATIONS FOR COMPLEMENTING THE MODEL IN THE FUTURE Further investigations can be envisaged lines would be making of the computational model and evaluating with real data the uses and statistical analysis performed by the OOSC ontology. To reach this computational model, quantitative parameters, observations and research objectives should be defined. An ontology - driven conceptual modeling of olive oil supply chain 70 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech Bibliography Antonucci, F. (2019) ‘A review on Blockchain applications in the agri‐food sector’, Journal of the Science of Food and Agriculture, 99(14), pp. 6129–6138. doi:10.1002/jsfa.9912. Beausang, C., Hall, C. and Toma, L. (2017). Food waste and losses in primary production: Qualitative insights from horticulture. Resources, Conservation & Recycling, 126: 177– 185 Bechhofer, S. (2009). OWL: Web Ontology Language. In: LIU, L., ÖZSU, M.T. (eds) Encyclopedia of Database Systems. Springer, Boston, MA. https://doi.org/10.1007/978-0-387-399409_1073 Beghin, J. C., A-C Disdier, S. Marette, and F. van Tongeren. (2011). “Measuring Costs And Benefits Of Non-Tariff Measures In Agri-Food Trade.” WP #110001, ISU Economics Working Paper Series. Iowa State University. Bouhaddane, Maria & Mili, Samir, (2018) "A Forecast of Internationalization Strategies for the Spanish Olive Oil Value Chain," 2018 International European Forum (163rd EAAE Seminar), February 5-9, 2018, Innsbruck-Igls, Austria 276854, International European Forum on System Dynamics and Innovation in Food Networks. https://doi.org/10.18461/pfsd.2018.%25x Carrington, D. (2022) ‘Spain and Portugal suffering driest climate for 1,200 years, research shows’, The Guardian, 4 July. Chandrasekaran, B., Josephson, J. R., & Benjamins, V. R. (1999). What are ontologies, and why do we need them? IEEE Intelligent Systems, 14(1), 20–26. http://doi.org/10.1109/5254.747902 Cranefield, S., & Purvis, M. (1999). UML as an ontology modelling language (Information Science Discussion Papers Series No. 99/01). University of Otago. Available at: http://hdl.handle.net/10523/932 Cummings, J. and Stacey, D. (2017) ‘The Mid Level Data Collection Ontology (DCO) - generic data collection using a mid level ontology’, Proceedings of the 9th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management [Preprint]. doi:10.5220/0006497501750182. DeGennaro, Ramon P., Market Imperfections. Journal of Financial Transformation, Vol. 14, 2005, FRB Atlanta Working Paper No. 2005-12, Available at SSRN: https://ssrn.com/abstract=775866 Dooley, D.M., Griffiths, E.J., Gosal, G.S. FoodOn: a harmonized food ontology to increase global food traceability, quality control and data integration. npj Sci Food 2, 23 (2018). https://doi.org/10.1038/s41538-018-0032-6 Fatemi Ghomi, S.M.T. and Asgarian, B. (2019), "Development of metaheuristics to solve a transportation inventory location routing problem considering lost sale for perishable goods", Journal of Modelling in Management, Vol. 14 No. 1, pp. 175-198. Fatemi Ghomi, S.M.T. and Asgarian, B. (2019), "Development of metaheuristics to solve a transportation inventory location routing problem considering lost sale for perishable An ontology - driven conceptual modeling of olive oil supply chain 71 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech goods", Journal of Modelling in Management, Vol. 14 No. 1, pp. 175198. https://doi.org/10.1108/JM2-05-2018-0064 Food loss and waste database: Technical platform on the measurement and reduction of food loss and waste: Food and Agriculture Organization of the United Nations. Retrieved November 15, 2022, from https://www.fao.org/platform-food-loss-waste/flw-data/en/ FoodOn (2022) A farm to fork ontology, Foodon. Available at: https://foodon.org/ Geerts, G. and O’Leary, D.E. (2014) ‘A supply chain of things: The Eaglet ontology for highly visible supply chains’, SSRN Electronic Journal [Preprint]. doi:10.2139/ssrn.2412741. Gogolla, M. (2009). Unified Modeling Language. In: LIU, L., ÖZSU, M.T. (eds) Encyclopedia of Database Systems. Springer, Boston, MA. https://doi.org/10.1007/978-0-387-399409_440 Gómez-Pérez, A. (2001), Evaluation of ontologies. Int. J. Intell. Syst., 16: 391409. https://doi.org/10.1002/1098-111X(200103)16:3<391::AID-INT1014>3.0.CO;2-2 Guido R., Mirabelli G.,Palermo E., Solina V., (2020). A framework for food traceability: case study - Italian extra-virgin olive oil supply chain. Int.J.Ind.Eng.Man. ISSN 2683-345X, vol.11,No.1,pp50-60. Guido, R. (2020) ‘A framework for food traceability: Case study – italian extra-virgin olive oil supply chain’, International Journal of Industrial Engineering and Management, 11(1), pp. 50–60. doi:10.24867/ijiem-2020-1-252. H. Sack. OpenHPI Tutorials. Semantic web technologies (2020) (Video*). https://www.youtube.com/@xhakil26/videos How can digital traceability increase trust in the agrifood industry? (2023) EIT Food. Available at: https://www.eitfood.eu/blog/how-can-digital-traceability-increase-trust-in-theagrifood-industry Ibañez, M. (2020) Tratamiento de efluentes de la industria del aceite de oliva mediante el ejemplo de hongos. thesis. Availabe at: TFM_MartaIbáñezLazaro.pdf (uniovi.es) Información de Consumo de Aceite (2022) Información de consumo de aceite. Available at: https://www.mapa.gob.es/es/alimentacion/temas/consumo-tendencias/panel-deconsumo-alimentario/ultimos-datos/consumo-aceite.aspx Jedermann, R., Nicometo, M., Uysal, I., & Lang, W. (2014). Reducing food losses by Intelligent Food Logistics. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 372(2017), 20130302. http://doi.org/10.1098/rsta.2013.0302 Jerôme Chandes, Dominique Estampe. Logistics Performance of Actors in the Wine Supply Chain. Supply Chain Forum: An International Journal, 2003, 4 (1), pp.12-27. ffhal-00767174 La Sexta (2022) ‘El aceite de oliva tendrá problemas de abastecimiento este año: habrá menos y será más caro’, 4 July. Available at: https://www.lasexta.com/noticias/nacional/aceiteoliva-tendra-problemas-abastecimiento-este-ano-habra-menos-sera-mascaro_20221005633de369cff9ff00013bd2a0.html Manning, L. (2015) ‘Determining value in the food supply chain’, British Food Journal, 117(11), pp. 2649–2663. doi:10.1108/bfj-02-2015-0049. An ontology - driven conceptual modeling of olive oil supply chain 72 Escola d’Enginyeria Agroalimentària i de Biosistemes de Barcelona UPC - BarcelonaTech MARM (2010) THE VALUE CHAIN AND PRICE FORMATION IN THE SPANISH OLIVE OIL INDUSTRY Available at: https://www.internationaloliveoil.org/wpcontent/uploads/2019/12/CADENADEVALOR-ENG.pdf Minh Hoang Lien Vo & Quang Hoang (2020) Transformation of UML class diagram into OWL Ontology, Journal of Information and Telecommunication, 4:1, 1-16, DOI: 10.1080/24751839.2019.1686681 Møller, A. (2017) The International Framework for Food Description, LanguaL. Available at: https://www.langual.org/default.asp Nathan, D. (2021) ‘From monopoly to monopsony capitalism’, The Indian Journal of Labour Economics, 64(4), pp. 843–866. doi:10.1007/s41027-021-00350-w. Nofer, M., Gomber, P., Hinz, O. Blockchain. Bus Inf Syst Eng 59, 183–187 (2017). https://doi.org/10.1007/s12599-017-0467-3 Nuijten, S. (2020) THE EIT FOOD TRUST REPORT. rep. TrustTracker. Available at: https://www.eitfood.eu/media/news-pdf/EIT_Food_Trust_Report_2020.pdf OECD Library (2022) Quarterly International Trade Statistics. Issue 3 ISSN: 23130857. doi:10.1787/23130857 Östergren, K., Gustavsson, J., Bos-Brouwers, H., Timmermans, T., Hansen, O.-J., Møller, H., Anderson, G., O’Connor, C., Soethoudt, H., Quested, T., Easteal, S., Politano, A., Bellettato, C., Canali, M., Falasconi, L., Gaiani, S., Vittuari, M., Schneider, F., Moates, G., Waldron, K., Redlingshöfer, B., 2014. FUSIONS Definitional Framework for Food Waste. FUSIONS. Pizzuti, T. (2017) ‘Mesco (Meat Supply Chain Ontology): An ontology for supporting traceability in the meat supply chain’, Food Control, 72, pp. 123–133. doi:10.1016/j.foodcont.2016.07.038. Ruben, R., Slingerland, M. and Nijhoff, H. (2006) Agro-food chains and networks for Development. Dordrecht, NE: Springer. Sakka, O., Millet, P.A., Botta-Genoulaz, V., 2010. An OWL based ontology of SCOR model: a prerequisite for strategic alignment. 8th Int. Conf. Model. Simulation-MOSIM’10 “Evaluation Optim. Innov. Prod. Syst. goods Serv. Sanchez, C. (2021) Datos sobre el monopolio de la Producción de semillas en el Mundo, Misión Verdad. Available at: https://misionverdad.com/datos-sobre-el-monopolio-de-laproduccion-de-semillas-en-el-mundo Smartdraw (1994-2023) UML diagram, UML Diagram - Everything You Need to Know About UML Diagrams. Available at: https://www.smartdraw.com/uml-diagram/ Staab, S. and Studer, R. (2009) Handbook on ontologies. 2nd edn. Dordrecht Netherlands: Springer. Sylvester, G. (2019) “E-agriculture in action: Blockchain for agriculture.” Bangkok: FAO. https://www.fao.org/3/CA2906EN/ca2906en.pdf Taylor, D.H. (2005), "Value chain analysis: an approach to supply chain improvement in agri‐food chains", International Journal of Physical Distribution & Logistics Management, Vol. 35 No. 10, pp. 744-761. https://doi.org/10.1108/09600030510634599