Hierarchical Path Planning for Multi-agent Systems Situated in Informed Virtual Geographic Environments

Multi-Agent Geo-Simulation (MAGS) is a modelling and simulation paradigm which involves a large number of autonomous situated agents evolving in, and interacting with, an explicit description of a geographic environment called a Virtual Geographic Environment (VGE). Path planning in MAGS has to be solved in real time, often under constraints of limited memory and CPU resources. Moreover, the computational cost of path planing increases in complex and large-scale VGEs. In addition, most current planners only provide agents with obstacle-free paths and do not take into account the environments' topologic and semantic characteristics nor the agents' capabilities. In this paper, we propose a novel approach to build a semantically-enhanced and geometrically-accurate VGE called an Informed VGE (IVGE). We also present a hierarchical path planning algorithm which takes advantage of this IVGE's rich description in order to provide autonomous situated agents with plausible paths with respect to both the environment'sand the agents' characteristics.