An Extended Evolutionary Learning Approach For Multiple Robot Path Planning In A Multi-Agent Environment

This paper describes an extended Genetic Algorithm Approach for path planning of multiple mobile robots with obstacle detection and avoidance in static and dynamic scenarios. Through the software Netlogo, used in simulations of multi-agent applications, a model was developed for the given problem. The model, which contains multiple robots and a scenario with several dynamic and static obstacles, is responsible for determining the best path used by the robots to achieve the goal state in a shorter number of steps and avoiding collisions. Additionally, a performance evaluation of this model in comparison with A* algorithm is presented.