Real Time Multi-agent Path Planning for Disaster Recovery, Case Study: RoboCup Rescue

Abstract

In this work an implementation prototype for GMAS is proposed. As a primary step, MAPP problem was addressed. Choosing this problem is derived from the fact that path planning consumes a significant part of the computation. The solution is applied on USAR domain. This implementation follows the design architecture of the RoboCup Rescue simulator which is considered the most famous benchmark for rescue systems implementation now. In USAR the scale of problems involves both a large number of agents and a large number of tasks that must be allocated to these agents. Besides, these tasks and their characteristics change over time. Little information about the whole scenario, if any, are available to all agents. Each agent has different capabilities and limited resources to perform its tasks. The problem is how to find, in a distributed fashion, an appropriate task allocation that represents the best match among agents and tasks. Implementing these systems requires a good knowledge in MAS development aspects related to MAS such as: MAS design, MAS coordination, task allocation, resources management and conflicts detection and resolution.