Multi-robot Path Planning Based on Cooperative Co-evolution and Adaptive CGA

Abstract

Multi-robot path planning is a challenge for mobile robots in AI. Multi-objective optimized algorithm based on cooperative co-evolution and CGA is brought up in this paper. Shortest path length, minimum time cost, smoothest and limited speed, obstacle-collide free and robot-collide free are the objectives and constraints to optimize. Linear combination of them is designed as evaluation function for CGA with self-adaptive crossover and mutation rate, combined with chaos disturbs. Finally 2D dynamic simulation has proved the efficiency of the algorithm.