Cognitive Territorial Agents for the Study of Urban Dynamics and Risks – ACTEUR

;

;

;

;

Every year, the number of urban residents is growing. Diverse questions related to sustainability are rise from this growth. For example, for large and attractive territories, which urban planning policies to implement? How to manage and prevent technological or environmental hazards? Decision makers have to take all of these issues into account when defining their urban planning policies. Unfortunately, the assessment of the impacts of possible policies is difficult due to the complex and stochastic interplay between society and infrastructure. One of the most promising approaches to face this difficulty is agent-based modeling. This approach consists in modeling the studied system as a collection of interacting decision-making entities called agents. An agent-based model can provide relevant information about the dynamics of the real-world urban system it represents. Moreover, it can allow to be used as a virtual laboratory to test new urban planning policies. The use of agent-based models to study urban systems is booming for the last ten years. Another tendency is the development of more and more realist models. However, if models have make a lot of progresses concerning the integration of geographical and statistical data, the agents used to represent the different actors influencing the dynamic of the system (inhabitants, decision makers...) are often simplistic (reactive agents). Yet, for some urban models, being able to integrate this cognitive agents, i.e. agents able to make complex reasoning such as planning to achieve their goals, is mandatory to improve the realism of models and test new scenarios. Unfortunately, developing large-scale models that integrates cognitive agents requires high-level programming skills. Indeed, if there are nowadays several software platforms that propose to help modelers to define their agent-based models through a dedicated modeling language (Netlogo, GAMA…) or through a graphical interface (Starlogo TNG, Modelling4All, Repast Symphony, MAGéo...), none of them are adapted to the development of such models by modelers with low level programming skills: either they are too complex to use (Repast, GAMA) or too limited (Netlogo, Starlogo TNG, Modelling4All, Repast Symphony, MAGéo). As a result, geographers and urban planners that have no programming skills have to rely on computer scientists to develop models, what slows the development and the use of complex and realist spatial agent-based models. The objective of the ACTEUR project is to develop to help modelers, in particular geographers and urban planners, to design and calibrate through a graphical language cognitive agents able to act in a complex spatial environment. The platform has also for ambition to be used as a support of model discussion -participatory modeling- between the different actors concerned by a model (geographers, sociologists, urban planners, decision makers, representatives…). These tools will be integrated in the GAMA platform that enables to build large-scale models with thousands of hundreds of agents and that was already used to develop models with cognitive agents. In order to illustrate the utility and the importance of the developed tools, we will use them on two case studies. The first concerns the urban evolution of La Réunion island. The second case study will focus on the adaption to industrial hazards in Rouen. These two case studies are part of funded projects carried out by partners of the ACTEUR project.