During the last decade, Multi-Agent Geo-Simulation (MAGS) has attracted a growing interest from researchers and practitioners to simulate various phenomena in a variety of domains such as traffic simulation, crowd simulation, urban dynamics, environment monitoring, and changes of land use and cover, to name a few. Such approaches are used to study various phenomena (i.e. car traffic, crowd behaviours, etc.) involving a large number of simulated actors (implemented as software agents) evolving in, and interacting with, an explicit spatial environment representation usually called Virtual Geographic Environment (VGE). 

A critical step towards the development of MAGS is the creation of a VGE, using appropriate representations of the geographic space and of the objects contained in it (also called ‘situated objects’) in order to efficiently support the agents’ reasoning and various situated behavioural activities. In order to yield realistic MAGS, a VGE must not only integrate geometric data, but also several semantic notions about specific areas (knowledge about location and associated usage such as the fact that side-walks are dedicated to pedestrian motion). Our specific aim is to create semantically-enhanced and geometrically accurate virtual geographic environments (also called IVGE for Informed VGE) populated with situated agents and objects. The IVGE is exploited by software agents for decision making and spatial reasoning purposes. Our current research work addresses this issue and proposes an automated approach for the generation of IVGE. Our approach is based on realistic data provided by Geographic Information Systems (GIS) and uses an exact space decomposition technique (using the Constrained Delaunay Triangulation (CDT) approach) which is mapped to an informed graph whose nodes represent convex cells and arcs represent connectivity relations. The semantic information extracted from the GIS data is represented using the Conceptual Graphs (CGs) formalism. CGs are fundamentally based on concepts and relations notions which allow to build a Semantic Concepts Hierarchy (SCH) that evolves the semantic description of the virtual environment. The informed graph is then topologically and semantically (using SCH) abstracted in order to build a hierarchical graph representation of the geographic environment whose nodes encompass geometric, topologic, as well as semantic information. Our objective is to create a hierarchical environment model with integrated knowledge adapted for the simulation of advanced situated behavioural activities. To conclude, this research work aims at shifting the MAGS to a next generation of enhanced geographic environments modelling for the support of advanced cognitive behaviours and spatial reasoning.

We are currently working on two research projects which leverage our IVGE model. The first project called Crowd-MAGS is conducted in collaboration with Defence Research Canada (Valcartier) and deals with the simulation of military operations in urban settings involving civilian crowds. In the context of a crowd control research project, we built an IVGE representing a part of Quebec City where the Summit of Americas was held in 2001. The simulation results show the importance of the terrain topology (building, places, road network, etc.) on agents’ behaviours dynamics (path planning, perception, and navigation). The second project is a water resources monitoring project which involves water level control, soil quality analysis, and flood risks analysis by means of sensor webs. A Sensor Web (SW) is a computer network of many, spatially distributed devices using sensors to monitor environmental conditions at different locations. These sensor nodes consist of sensing, data sharing and processing, and communicating components. This project aims at improving the information that scientists and resource managers use to understand current environmental conditions within a watershed. In such a project we need sensors that collaborate to collect/filter data about the current situation, to adapt to spatial and physical changes, to warn users and help them make decisions. Providing sensors with an informed view of the geographic environment in which they are situated enables them to coordinate and to adapt to the environment changes (spatial reasoning using the semantic information associated with the IVGE and its dynamism thanks to the Geo-Simulation).