Mobile Robot Path Planning Based on Multi-objective Evolutionary Programming

Abstract

A new evolutionary programming algorithm for multi-objective optimization is proposed. It employs two elitism mechanisms and does not include fitness assignment. The proposed algorithm was applied to the mobile robot path planning with three objectives. The random approach combined with a heuristic method based on domain knowledge were used in the initialization process to motivate convergence, and three evolutionary operators named deletion, repair and smooth were designed to improve the searching efficiency of the algorithm. Simulation results indicate that the proposed algorithm can find a set of Pareto optimal solutions efficiently in a single run, from which the decision maker can make a choice.