Content uploaded by Patrick Taillandier
Author content
All content in this area was uploaded by Patrick Taillandier
Content may be subject to copyright.
A preview of the PDF is not available
A new BDI agent architecture based on the belief theory. Application to the modelling of cropping plan decision-making
Abstract and Figures
Agent-based simulations are now widely used to study complex systems. However, the problem of the agent design is still an open issue, especially for social ecological models, where some of the agents represent human beings. In fact, designing complex agents able to act in a believable way is a difficult task, in particular when their behaviour is led by many conflicting needs and desires. A widely used way to formalise the internal architecture of such complex agents is the BDI (Belief Desire-Intention) paradigm. This paradigm allows to design expressive and realistic agents, yet, it is rarely used in simulation context. A reason is that most agent architectures based on the BDI paradigm are complex to understand by non-computer-scientists. Moreover, they are often very time-consuming in terms of computation. In this paper, we propose a new architecture based on the BDI paradigm that copes with these two issues. In our architecture, the choice of the most relevant action by an agent is based on the belief theory. We present an application of our agent architecture to an actual model dedicated to cropping plan decision-making. This application that takes into plays thousands of farmer agents shows promising results.
Figures - uploaded by Patrick Taillandier
Author content
All figure content in this area was uploaded by Patrick Taillandier
Content may be subject to copyright.
... Liang et al., 2016;Muto et al., 2020;Truong et al., 2015) and crop selection decisions (e.g. Dury et al., 2010;Robert, Dury et al., 2016;Taillandier et al., 2012). These land-use and crop selection decisions are taken at a strategic or tactical level. ...
... Apart from two papers, these works did not explicitly represent uncertainty in their models. Taillandier et al. (2012)'s work is an exception here: their view was that farmers' desires were not mutually exclusive. They used a multi-criteria approach to selecting the farmers' desires. ...
... To our knowledge, no one has proposed a BDI agent architecture representing the operational decisions of a sugarcane grower. Also, apart from Taillandier et al. (2012) and Muto et al. (2020), to our knowledge, no one has proposed a farmer BDI agent which reasons with uncertainty. ...
Building computational models of agents in dynamic, partially observable and stochastic environments is challenging. We propose a cognitive computational model of sugarcane growers' daily decision-making to examine sugarcane supply chain complexities. Growers make decisions based on uncertain weather forecasts; cane dryness ; unforeseen emergencies; and the mill's unexpected call for delivery of a different amount of cane. The Belief-Desire-Intention (BDI) architecture has been used to model cognitive agents in many domains, including agriculture. However, typical implementations of this architecture have represented beliefs symbolically, so uncertain beliefs are usually not catered for. Here we show that a BDI architecture, enhanced with a dynamic decision network (DDN), suitably models sugarcane grower agents' repeated daily decisions. Using two complex scenarios, we demonstrate that the agent selects the appropriate intention, and suggests how the grower should act adaptively and proactively to achieve his goals. In addition, we provide a mapping for using a DDN in a BDI architecture. This architecture can be used for modelling sugarcane grower agents in an agent-based simulation. The mapping of the DDN's use in the BDI architecture enables this work to be applied to other domains for modelling agents' repeated decisions in partially observable, stochastic and dynamic environments.
... Some models propose to integrate both the definition of the objectives and of the crop allocation in the same process. For example, [13] propose to formulate the choice of crop rotation as a multi-criteria decision problem: for each plot, the farmer chooses a crop rotation, according to an evaluation at the farm level of several criteria (financial risks, expected incomes, workload and farmer habits). Other works focus only on the crop allocation problem. ...
... In this paper, the challenge is to simulate the behavior of the actors (positive approach) and not to find an optimal configuration, we thus chose to take inspiration from the work of [17], but, in order to be able to run simulation at a mediumscale, we used for the choice of crops a multi-criteria decision approach as presented in [13]. ...
... In this model, and contrarily to what has been done in [13], [17], an important choice has been made not to take the plots as spatial unit, but to divide the space into uniform cells. This choice results from two strong constraints: the absence of precise data on the plots and the computation time. ...
The rice-shrimp farming system is considered as a sustainable and beneficial model for the environment. However, the area of rice-shrimp was increasingly narrowed due to the trend of converting from rice to aquaculture by economic reasons. This paper aims to propose a medium scale land use change model for understanding the land use decision of farmers in adaptation to the environment and climate change. The model integrates a land-use decision making process based on multi-criteria selection where the main factors are land suitability, land convertibility, land use situation of neighbors, and profitability of land use patterns. Concerning the land use data, we used historical land use map in 2005, 2015 and 2019. Shrimp cultivation regions was completed by Landsat satellite image processing. The model has been calibrated by rice-shrimp map in 2015 and has been verified with the rice – shrimp map in 2019 of the My Xuyen district, Soc Trang province, Vietnam. The simulated results show that the rice-shrimp area was increasingly narrowed and has been converted to aquaculture land. In addition, the model tends to show that in a scenario of sea level rise of 15 cm in 2030, the share of rice-shrimp and shrimp tends to rise sharply, which is an important lesson for developing complex adaptive strategies of farmers.
... Về mặt ứng dụng mô hình hóa, Taillandier et al. (2012) đã đề xuất xây dựng thay đổi kiểu sử dụng đất như một bài toán phân tích đa tiêu chí. Đối với mỗi ô đất trên bản đồ, người nông dân chọn thay đổi cây trồng theo đánh giá ở cấp độ trang trại theo nhiều tiêu chí về rủi ro tài chính, dự kiến thu nhập, khối lượng công việc và thói quen của nông dân. ...
... Nghiên cứu đã xác định cho việc lựa chọn kiểu sử dụng đất dựa trên đánh giá đa tiêu chí theo cách tiếp cận của Taillandier et al. (2012). Các tiêu chí được chọn được chọn lọc từ các tiêu chí ảnh hưởng đến sự thay đổi sử dụng đất đã phân tích (Lambin, 2007;Phạm Thanh Vũ và ctv, 2013) cho trường hợp nghiên cứu gồm 4 tiêu chí: ...
Changes in using agricultural land pose challenges for land managers in terms of ensuring the implementation of local land-use plans. The paper aims to build a land-use change model for simulating land-use changes under the impacts of socioeconomic and environmental factors. The model was built based on Agent-based modeling approach using GAMA software. In which the land-use decision-making process is based on the multi-criteria selection that the main factors were land suitability, land convertibility, land-use situation of neighbors, and profitability of land-use patterns. The input data for simulation were the land-use maps in 2010, 2015 and 2020 of Tran De district, Soc Trang province. A new model was built and has been calibrated using land-use map in 2015 (with Kappa = 0.71). The model has been verified with the land use map in 2020 with the simulation error percentage, nRMSE, which was 5.2%. The simulated results showed that rice land tends to change to rice-vegetable, specialized crops and perennial fruit trees to respond to climate in 2030. The case study showed that the model is an essential tool for helping land managers and farmers to build adaptive land-use plans.
... For annual crops, landscapes change every year in relation with crop successions and rotation rules. Taillandier et al. (2012) formulated the choice of crop rotation as a multi-criteria decision problem: for each plot, the farmer chooses a crop rotation, following evaluations at the farm level based on several criteria (financial risk, expected income, workload, farmer's habits). Other works only focus on the crop allocation problem. ...
... As this algorithm is stochastic, different solutions of spatial crop allocation can be obtained between two executions depending on random choices for the construction of graph-edges. Unlike Taillandier et al. (2012), labor tasks were not explicitly evaluated, but plowing, sowing and hoeing dates were stochastically distributed in the computed crop successions to generate "noise" in calendars. ...
The present article proposes a suite of programs aimed at (i) representing landscape organization in relation with the functioning of annual crops or breeding systems, (ii) evaluating water movements from biomass and hydrological exchanges, and (iii) analyzing the transfer dynamics of nutrients, suspended matter or fecal bacteria based on particle tracking methods. Simulations provided indications about the underlying processes that drive exchanges and sink-source effects operating at the landscape scale. OPALE was tested in three agricultural contexts and biophysical situations. Water flows and flows of associated matters were compared to data recorded at the catchment outlet, and showed the efficiency of the algorithms developed in the generic OPALE libraries. This article describes the underlying hypotheses and the full mathematical framework and procedures used to assess landscape eco-functionality. Several examples are given to illustrate the use of OPALE in landscape reconfiguration prospects (e.g. influence of landscape composition and structure on plant transpiration, stream flow, erosion, nutrient and organism fluxes in water) for the agroecological transition.
... The cropping plan decision module simulates allocation of cropping system (combination of rotation and crop management strategies) in each arable field of the farm. The choice of a rotation is carried out through a multi-criteria decision-making method based on Dempster-Shafer belief theory included in a Belief-Desire-Intention architecture (Taillandier et al., 2012). This multi-criteria method allows simulation of a decision made with incompleteness, uncertainly, and imprecision of knowledge. ...
... Notre démarche de travail repose sur plusieurs constructions, avec dans un premier temps la collecte de savoirs locaux (par enquêtes, visites de terrain et ateliers d'échanges) ou construits portant sur les déterminants du paysage, avec en particulier l'analyse de travaux préalables sur les contraintes et modes d'utilisation de l'espace dans les systèmes d'élevage (Marie et al., 2016) ou les déterminants des choix d'utilisation de l'espace et de l'évolution des surfaces dans le cas du maraîchage et des grandes cultures (Taillandier et al., 2012). On a donc eu pour souci de développer un module de calcul générique -la fabrique du paysage -pour aborder un ensemble vaste de questionnements soulevés dans des contextes très différents de bassins de production agricole. ...
... Instead of using mathematical approaches, complex agents can be designed based on the Belief-Desire-Intention (BDI) paradigm. It is developed on cognitive frameworks and embeds beliefs, desires, and intentions into individual agents [80]. e BDI architecture regarding the decisionmaking mechanism indicated that the decision-making behaviors of agents, such as the perception of information towards the current system state, desirable and motivational state, and final act of agents, greatly influenced the model output [81]. ...
This article presents an overview of the agent-based modeling and simulation approach and its recent developments in transport fields, with the purpose of discovering the advantages and gaps and encouraging more valuable investigations and applications of agent-based models. We clarify the agent-based model from agents, the background of development, and the basic structure applied in transport systems. Then, the agent-based transport modeling toolkits are discussed. The applications of agent-based models in transport systems are reviewed in three time scale models followed by an additional discussion of hybrid modeling approaches. The extensive modeling of the beliefs, desires, learning, and adaptability of individuals and the optimization problems using agent-based models are explored. Besides, we point out some limitations in terms of calibration and validation procedure, agents’ behavior modeling, and computing efficiency. In conclusion, some recommendations are given and suggest potential and insightful directions such as Big Data and Digital Twin for future research.
... (BDI ) agent architecture, as done for operational decision-making by Martin-Clouaire (2017) and for structural decisions in MAELIA. In MAELIA, each farmer agent establishes annually a cropping plan through rotations that are chosen according to a set of criteria based on a BDI architecture (Taillandier et al., 2012). This allows the farmer agent to make a decision even in the absence of complete information, based on four criteria: profit maximization, similarity to the last cropping plan, minimization of financial risks and workload. ...
In a context of global change, management of water resources between actors with sometimes conflicting objectives should be revisited. Modeling and simulation tools, dedicated to prospective analysis, are being developed at Irstea, in partnership with stakehoders of water management. Hydrological processes are rather well know and modeling and simulation platforms are available for their simulation. However, before they can produce relevant prospective scenarios, several improvments are required, in particular accounting for feedbacks loops between social and physical dynamics. Distributed hydrological modeling enables the simulation of hydrological dynamics considering spatial heterogeneity of the catchments and consequently gives the opportunity to imagine local feedback loops related to human activites. The objectives of the thesis is to couple an existing distributed hydrological model to an agent based model representing social dynamics and driving the local feedback loops related to human activities. We propose to use the theory of situed action (providing a framework to describe individual actions and decisions) and the concept of stigmergy (providing a framework to describe indirect interactions amongst actors) to build the agent based model. In this context, the decision making process of one or several agents is based on the set of actions that he can possibly realize at each moment. this implies to put the focus on the variables of interest of the agent (perceived variables and control variables, including hydrological variables) that are most of the time local variables. Hence, the exercice of coupling the agent model and the hydrological model will consist mainly in defining agents's variables of interest and to relate them to state variables or outputs of the hydrological model. This link will enable to specify local feedback loops related to human activities and furthermore to get spatial simulation outputs matching stakeholders' variables of interest. The scope of the PhD thesis is limited to the main sectors of activity of the two chosen sub-catchments of the Durance river. We will iteratively measure the impact of the implemented local feedback loops on the outputs computed at the scale of the sub-catchment. The impacts of local and sectoral adaptations (mainly from the irrigated agriculture sector constituting the main water use), on the global behavior oh the hydrosystem, under constraints of management strategies specified at the scale of the sub-catchment (incentives, regulations,..).
... Two important developments might foster a greater recognition of the importance of robust cognitive models for some applications. First, efforts are being made to develop cognitive frameworks within modeling packages, such as the BDI framework in MATSim (Horni, Nagel, and Axhausen 2016) or (Taillandier, Therond, and Gaudou 2012). Second, frameworks such as ODD + D (Müller et al. 2013) and Modeling Human Behavior (Schlüter et al. 2017) which attempt to provide a means for communicating and comparing different theories of individual human decision-making are showing promise. ...
Despite reaching a point of acceptance as a research tool across the geographical and social sciences, there remain significant methodological challenges for agent‐based models. These include recognizing and simulating emergent phenomena, agent representation, construction of behavioral rules, and calibration and validation. While advances in individual‐level data and computing power have opened up new research avenues, they have also brought with them a new set of challenges. This article reviews some of the challenges that the field has faced, the opportunities available to advance the state‐of‐the‐art, and the outlook for the field over the next decade. We argue that although agent‐based models continue to have enormous promise as a means of developing dynamic spatial simulations, the field needs to fully embrace the potential offered by approaches from machine learning to allow us to fully broaden and deepen our understanding of geographical systems.
The agent-based simulation is a powerful tool to study complex systems. It allows to take into account different levels of granularity, as well as the heterogeneity of the entities composing the system. One of the main issues raised by these simulations concerns the design of the agent behavior. Indeed, when the agent behavior is led by many conflicting needs, desires and necessities, its definition can be very complex. In order to address this issue, we propose a new formalism to define the agent decision making process. This formalism is based on the belief theory, which is a formal theory about reasoning that allows to manage information incompleteness, uncertainty and imprecision. An application of the approach is proposed in the context of a model dedicated to cropping plan decision-making.
In the wake of today's drive for increasingly realistic video game environments, the credibility of non-player characters (NPC) behavior has become an important issue. This paper describes MHiCS (Motivational and Hierarchical with Classifier Systems), a control architecture for action selection that can be embedded in such characters. It implements mechanisms stemming from the animat approach, which aims at conceiving autonomous and adaptable entities able to survive in unpredictable environments. The originality of MHiCS lies in its combination of a motivation mechanism with intelligible condition-action rules (Classifiers Systems) and online unsupervised learning processes. It has been successfully tested in the Half Life game, against efficient hand-tuned NPC. This work contributes to the design of Classifiers Systems operating in non-deterministic environments and to the improvement of decisional autonomy in non-player characters.
Smets, P. and R. Kennes, The transferable belief model, Artificial Intelligence 66 (1994) 191–234.
We describe the transferable belief model, a model for representing quantified beliefs based on belief functions. Beliefs can be held at two levels: (1) a credal level where beliefs are entertained and quanti?ed by belief functions, (2) a pignistic level where beliefs can be used to make decisions and are quantified by probability functions. The relation between the belief function and the probability function when decisions must be made is derived and justified. Four paradigms are analyzed in order to compare Bayesian, upper and lower probability, and the transferable belief approaches.
Simulation models that describe autonomous individual organisms (individual based models, IBM) or agents (agent-based models, ABM) have become a widely used tool, not only in ecology, but also in many other disciplines dealing with complex systems made up of autonomous entities. However, there is no standard protocol for describing such simulation models, which can make them difficult to understand and to duplicate. This paper presents a proposed standard protocol, ODD, for describing IBMs and ABMs, developed and tested by 28 modellers who cover a wide range of fields within ecology. This protocol consists of three blocks (Overview, Design concepts, and Details), which are subdivided into seven elements: Purpose, State variables and scales, Process overview and scheduling, Design concepts, Initialization, Input, and Submodels. We explain which aspects of a model should be described in each element, and we present an example to illustrate the protocol in use. In addition, 19 examples are available in an Online Appendix. We consider ODD as a first step for establishing a more detailed common format of the description of IBMs and ABMs. Once initiated, the protocol will hopefully evolve as it becomes used by a sufficiently large proportion of modellers.
Multi-agent simulations are powerful tools to study complex systems. However, a major difficulty raised by these simulations concerns the design of the agent behavior. Indeed, when the agent behavior is lead by many conflicting criteria (needs and desires), its definition is very complex. In order to address this issue, we propose to use the belief theory to formalize the agent behavior. This formal theory allows to manage the criteria incompleteness, uncertainty and imprecision. The formalism proposed divides the decision making process in three steps: the first one consists in computing the basic belief masses of each criterion; the second one in merging these belief masses; and the last one in making a decision from the merged belief masses. An application of the approach is proposed in the context of a model dedicated to the study of the avian flu propagation.
A multivalued mapping from a space X to a space S carries a probability measure defined over subsets of X into a system of upper and lower probabilities over subsets of S. Some basic properties of such systems are explored in Sects. 1 and 2. Other approaches to upper and lower probabilities are possible and some of these are related to the present approach in Sect. 3. A distinctive feature of the present approach is a rule for conditioning, or more generally, a rule for combining sources of information, as discussed in Sects. 4 and 5. Finally, the context in statistical inference from which the present theory arose is sketched briefly in Sect. 6.
Jason is an Open Source interpreter for an extended version of AgentSpeak - a logic-based agent-oriented programming language - written in Java™. It enables users to build complex multi-agent systems that are capable of operating in environments previously considered too unpredictable for computers to handle. Jason is easily customisable and is suitable for the implementation of reactive planning systems according to the Belief-Desire-Intention (BDI) architecture. Programming Multi-Agent Systems in AgentSpeak using Jason provides a brief introduction to multi-agent systems and the BDI agent architecture on which AgentSpeak is based. The authors explain Jason's AgentSpeak variant and provide a comprehensive, practical guide to using Jason to program multi-agent systems. Some of the examples include diagrams generated using an agent-oriented software engineering methodology particularly suited for implementation using BDI-based programming languages. The authors also give guidance on good programming style with AgentSpeak. Programming Multi-Agent Systems in AgentSpeak using Jason. Describes and explains in detail the AgentSpeak extension interpreted by Jason and shows how to create multi-agent systems using the Jason platform. Reinforces learning with examples, problems, and illustrations. Includes two case studies which demonstrate the use of Jason in practice. Features an accompanying website that provides further learning resources including sample code, exercises, and slides. This essential guide to AgentSpeak and Jason will be invaluable to senior undergraduate and postgraduate students studying multi-agent systems. The book will also be of interest to software engineers, designers, developers, and programmers interested in multi-agent systems.
The purpose of the method presented in this paper is to assess additive utility functions which aggregate multiple criteria in a composite criterion, using the information given by a subjective ranking on a set of stimuli or actions (weak-order comparison judgments) and the multicriteria evaluations of these actions. It is an ordinal regression method using linear programming to estimate the parameters of the utility function.Stability and sensitivity analysis leads to the assessment of a set of utility functions by means of post-optimality analysis techniques in linear programming.Finally, a simple illustrative example is presented and some extensions of the method are proposed.