具有修正策略的改进NSGA-Ⅱ三维路径规划

The Three-Dimensional Path Planning Based on an Improved NSGA2-Ⅱ with Modified Strategy

针对传统多目标遗传算法存在收敛速度慢和难以得到Pareto最优解的缺点,提出了一种在三维环境下具有修正策略的改进带精英策略的非支配排序的遗传算法(NSGA-Ⅱ)。首先建立能使路径最短、能耗最小、起伏最少的多目标函数;其次加入修正算子来减少冗余的路径点,实现快速收敛;然后在选择算子中加入辅助决策算子来比较优先级,提高解的多样性。为了测试改进算法的效果,将传统算法与改进算法进行对比,改进算法得到的解更优且在不同环境下具有多个Pareto前沿分布解,其中修正算子使迭代次数减少了约63%,验证了改进算法的可行性和有效性。

In view of the shortcomings of convergence speed and difficult to get Pareto’s optimal solution,an improved NSGA-Ⅱ with a modified strategy is proposed to plan the optimal collision-free path for a mobile robot in three-dimensional environment.Firstly,the optimization targets involving the shortest path,minimum energy consumption and least fluctuation were determined.Secondly,a modified operator was added to reduce redundant path points to achieve fast convergence.Meanwhile,an auxiliary decision operator was added to the selection operator to compare the priorities and reinforce the diversity of solutions.The traditional algorithm compares with the improved algorithm in order to verify the effectiveness of the improved algorithm,using improved NAGA-Ⅱ not only obtains a more satisfactory Pareto solution,but also get better diversity of solutions in Pareto optimal front in different cases,the modified operator reduces the number of iterations by about 63%,which verify that the improved algorithm owns better feasibility and effectiveness.