A combined coordination technique for Multi-Agent Path Planning

Multi-Agent Path Planning (MAPP) can be considered the basic building block for implementing a reliable Multi-Agent Systems capable for interacting with real world. MAPP main task is how to get the appropriate path for movable agents where each agent is considered a dynamic obstacle to the others. Planning the exact path for each movable agent in a highly dynamic environment is a difficult task. This paper presents a combined coordination technique for solving MAPP problem; this technique takes the benefits of both agents´ capabilities and system resources when available. Moreover it combines three planning algorithms, namely; task allocation using extended global assignment problem model, Checkpoint Priority Action Database algorithm for MAPP coordination and Potential Fields Methods for local path planning.