Sébastien Picault

IODA: an interaction-oriented approach for multi-agent based simulations

Multi-Agent Systems (MAS) design methodologies and Integrated Development Environments exhibit many interesting properties that also support simulation design. Yet, in their current form, they are not appropriate enough to model Multi-Agent Based Simulations (MABS). Indeed, their design is focused on the functionalities to be achieved by the MAS and the allocation of these functionalities among software agents. In that context, the most important point of design is the organization of the agents and how they communicate with each other. On the opposite, MABS aim at studying emergent phenomena, the origin of which lies in the interactions between entities and their interaction with the environment. In that context, the interactions are not limited to exchanging messages but can also be fundamental physical interactions or any other actions involving simultaneously the environment and one or several agents. To deal with this issue, this paper presents the core notions of the Interaction-Oriented Design of Agent simulations (IODA) approach to simulation design. It includes a design methodology, a model, an architecture and also JEDI, a simple implementation of IODA concepts for reactive agents. First of all, our approach focuses on the design of an agent-independent specification of behaviors, called interactions. These interactions are not limited to the analysis phase of simulation: they are made concrete both in the model and at the implementation stage. In addition, no distinction is made between agents and objects: all entities of the simulation are agents. Owing to this principle, designing which interactions occur between agents, as well as how agents act, is achieved by means of an intuitive plug-and-play process, where interaction abilities are distributed among the agents. Besides, the guidelines provided by IODA are not limited to the specification of the model as they help the designer from the very beginning towards a concrete implementation of the simulation.

An interaction-oriented model for multi-scale simulation

The design of multiagent simulations devoted to complex systems, addresses the issue of modeling behaviors that are involved at different space, time, behavior scales, each one being relevant so as to represent a feature of the phenomenon. We propose here a generic formalismintended to represent multiple environments, endowed with their own spatiotemporal scales and with behavioral rules for the agents they contain. An environment can be nested inside any agent, which itself is situated in one or more environments. This leads to a lattice decomposition of the global system, which appears to be necessary for an accurate design of multi-scale systems. This uniform representation of entities and behaviors at each abstraction level relies upon an interaction-oriented approach for the design of agent simulations, which clearly separates agents from interactions, from the modeling to the code. We also explain the implementation of our formalism within an existing interaction-based platform.

The MICRobES Project, an Experimental Approach towards “Open Collective Robotics”

We focus here on “Open Collective Robotics”, a research domain interested in studying concepts and techniques which could lead to true collectivities between humans and robots. We also present the MICRobES project1, a long-term experiment intended to investigate how to design such collectivities. Therefore, the MICRobES robots have no prior specific functionality, they have to develop individual competences, a social organization, and sociability towards humans in order to “survive” in a dynamic environment.

Modeling Adaptive Multi-Agent Systems Inspired by Developmental Biology

This paper addresses the issue of adaptive multi-agent systems and their design based on living systems features such as phylogeny and ontogeny. We argue that the evolutionary design of agents behaviors implies several specific features that are missing in classical evolutionary approaches. Therefore we propose a new approach that would be more adequate to MAS, and present a model for building MAS as the result of evolving, interacting, self-organizing agents. We finally mention a use of such an approach for the embodiment of robots.

How to Avoid Biases in Reactive Simulations

In order to ensure simulations reproducibility, particular attention must be payed to the specification of its model. This requires adequate design methodologies, that enlightens modelers on possible implementation ambiguities – and biases – their model might have. Yet, because of not adapted knowledge representation, current reactive simulation design methodologies lack specifications concerning interaction selection, especially in stochastic behaviors. Thanks to the interaction-oriented methodology IODA – which knowledge representation is fit to handle such problems – this paper provides simple guidelines to describe interaction selection. These guidelines use a subsumption like-structure, and focus the design of interaction selection on two points : how the selection takes place – for instance first select the interaction, and then select the partner of the interaction, or first a partner and then an interaction – and the nature of each selection – for instance at random, or with a utility function. This provides a valuable communication support between modelers and computer scientists, that makes the interpretation of the model and its implementation clearer, and the identification of ambiguities and biases easier.

Further Comparison between ATNoSFERES and XCSM

In this paper we present ATNoSFERES, a new framework based on an indirect encoding Genetic Algorithm which builds finite-state automata controllers able to deal with perceptual aliazing. In the context of our ongoing line of research, we compare it with XCSM, a memory-based extension of the most studied Learning Classifier System, XCS, through two benchmark experiments. We focus in particular on internal state generalization, and add special purpose features to ATNoSFERES to fulfill that comparison. We then discuss the role played by internal state generalization in the experiments studied.

Virtual Customers in an Agent World

The relevance of multi-agent systems (MAS) has been demonstrated in computer simulations or video games where many autonomous entities interact in a complex and dynamic environment. Serious Games (SG) are a new discipline situated at the edge of computer simulation and games. We advocate that a certain category of SG, where the player is immersed in a 3d environment, represents a particularly interesting testbed for MAS, for they introduce novel and inspirational problematics for the community. In this paper, we explore the challenges posed by these immersive SG to the MAS approach. Particularly, we demonstrate that the IODA interaction-oriented MAS approach, answers these new problematics with efficacy. We illustrate our discussion with a SG project developed in our team.

From Real Purchase to Realistic Populations of Simulated Customers

The use of multiagent-based simulations in marketing is quite recent, but is growing quickly as the result of the ability of such modeling methods to provide not only forecasts, but also a deep understanding of complex interactions that account for purchase decisions. However, the confidence in simulation predictions and explanations is also tightly dependent on the ability of the model to integrate statistical knowledge and the situatedness of a retail store. In this paper, we propose a method for automatically retrieving prototypes of consumer behaviors from statistical measures based on real data (receipts). After preliminary experiments to validate this data mining process, we show how to populate a multiagent simulation with realistic agents, by initializing some of their goals with those prototypes. Endowed with the same overall behavior, validated in earlier experiments, those agents are put into a spatially realistic store. During the simulation, their actual actions reflect the diversity of real customers, and finally their purchase reproduce the original clusters. Besides, we explain how such statistically realistic simulation may be used to support decision in retail, and be extended to other application domains.

An experimental comparison between ATNoSFERES and ACS

After two papers comparing ATNoSFERES with XCSM, a Learning Classifier System with internal states, this paper is devoted to a comparison between ATNoSFERES and ACS (an Anticipatory Learning Classifier System). As previously, we focus on the way perceptual aliazing problems encountered in non-Markov environments are solved with both kinds of systems. We shortly present ATNoSFERES, a framework based on an indirect encoding Genetic Algorithm which builds finite-state automata controllers, and we compare it with ACS through two benchmark experiments. The comparison shows that the difference in performance between both system depends on the environment. This raises a discussion of the adequacy of both adaptive mechanisms to particular subclasses of non-Markov problems. Furthermore, since ACS converges much faster than ATNoSFERES, we discuss the need to introduce learning capabilities in our model. As a conclusion, we advocate for the need of more experimental comparisons between different systems in the Learning Classifier System community.

A Multi-Level Multi-Agent Simulation Framework in Animal Epidemiology

In order to recommend better control measures in public or animal health, epidemiologists incorporate ever-finer details in their models, from individual diversity to public policies, which often involve several observation scales. Due to the variety of modelling paradigms, it becomes more and more difficult to compare hypotheses and outcomes, all the more that the increased complexity of simulation programs is not yet counterbalanced by design principles nor by software engineering methods. We propose in this paper to use the multi-level agent-based paradigm to integrate existing methods within a common interface, provide a separation between concerns and reduce the part of code devoted to model designers. We illustrate our approach with an application to the Q fever disease in cattle.