摘要： 针对船舶全局路径规划中的多路径规划问题，提出一种基于小生境粒子群优化的多路径规划方法。借鉴遗传算法求解多峰函数的思想，在基本粒子群优化算法的基础上，引入小生境生成、隔离进化和交叉算子等策略，分别采用主群粒子进行区域划分，使用子群粒子完成区域内的局部寻优，从而实现船舶多路径规划。仿真结果表明，该方法能正确、有效地规划出多条最优及次优路径。

关键词: 多路径规划, 粒子群优化, 小生境, 隔离进化, 交叉算子, 多种群

Abstract: For the problem of multi-path planning in global path planning for ships, a multi-path planning method based on niche Particle Swarm Optimization(PSO) is proposed. Based on the idea of Genetic Algorithm(GA) to solve multi-modal function, it takes the formation of niche, isolated evolution and crossover operator into the PSO, uses the main particle swarm to decompose the region, and the sub-particle swarm to optimize the local region respectively to accomplish the multi-path planning for ships. Simulation results show that this method can plan multiple optimal and sub-optimal paths correctly and efficiently.

Key words: multi-path planning, Particle Swarm Optimization(PSO), niche, isolated evolution, crossover operator, multi-population

中图分类号: 

• TP391