多目标遗传算法在机器人路径规划中的应用

针对具有多个优化目标的机器人全局路径规划问题，提出一种改进的多目标优化遗传算法。在初始群体的生成中，采用把随机法和基于问题先验知识的启发式方法相结合的策略，以加快收敛速度；在遗传算子的设计中，引入删除、修复和平滑算子，以提高算法的搜索效率；在选择算子中。加入避免外部存储器中出现相同个体的机制，以防止早熟收敛。仿真结果表明：该文算法运行一次能够有效地产生一组近似Pareto最优路径解。

Application of Multi-objective Optimization Genetic Algorithm to Robot Path Planning

An improved multi-objective optimization genetic algorithm is proposed to optimize the problem of robot global path planning with multiple objectives. The random approach combined with the heuristic method based on domain knowledge are employed in the initialization to motivate the convergence speed and three genetic operators named deletion, repair and smooth are adopted to improve the searching efficiency of the algorithm. In the selection operator, a strategy that avoids the appearance of the same individuals in the archive is incorporated to prevent premature. Simulation results indicate that the proposed algorithm can find a set of approximate Pareto optimal solutions efficiently in one run.