摘要
为了解决标准粒子群算法在无人机三维路径规划中存在的易陷入局部最优、动态化不足和路径平滑性差等问题,提出了一种基于粒子群算法和遗传算法的,融入K均值精英化和柯西变异的优化算法;采用K均值聚类算法进行精英初始化,优化粒子种群的分布;动态化学习因子,强化惯性权重的全局性,保留粒子群算法收敛速度快的优点;融入遗传思想,采用柯西变异的方法,提高寻解最优解的能力;在对比实验中,模拟了实际的复杂三维环境,选取了路径总长度、飞行高度差以及马尔科夫生存状态组成目标函数;结果表明改进算法的鲁棒性提高了98%,求解质量相较于IG-PSO算法和IC-PSO算法分别提高了5.8%和10.6%,验证了优化后方法的有效性和鲁棒性。
In order to solve the problems of the standard particle swarm algorithm in UAV 3D path planning,such as easy to fall into local optimum,insufficient dynamics and poor path smoothing,an optimization algorithm based on particle swarm algorithm and genetic algorithm,incorporating K-mean elitism and Cauchy variation is proposed.K-mean algorithm is used to perform the elite initialization and optimize the distribution of particle populations,dynamic learning factor,strengthening the global feature of inertial weights,keeping the advantage of fast convergence pf particle swarm optimization algorithm,integrating the genetic thinking,the Cauchy mutation method improves the ability to seek the optimal solution.In the comparison experiments,an actual complex 3D environment is simulated,and the total path length,flight altitude difference and Markovian survival state are selected as an objective function;the results show that the robustness of the improved algorithm is improved by 98%,and the solution quality of the improved algorithm is higher than that of the IG-PSO algorithm and IC-PSO algorithm by 5.8%and 10.6%,respectively,verifying the effectiveness and robustness of the optimized method.
作者
于力涵
洪儒
吴宇伦
谢迎娟
YU Lihan;HONG Ru;WU Yulun;XIE Yingjuan(School of Information Science and Engineering,Hohai University,Changzhou 213002,China)
出处
《计算机测量与控制》
2023年第8期259-266,共8页
Computer Measurement &Control
基金
国家自然科学基金(61573128)
国家自然科学基金(61701169)
国家重点研发计划(2018YFC0407101)
教育部产学合作协同育人项目(220803494162012,220603632072407)
国家级大学生创业训练项目(202210294232E)
江苏省大学生创新创业训练计划项目(202210294220Y)
河海大学本科实践教学改革研究项目(河海教务〔2022〕47号)
河海大学创新性实验项目(河海教务〔2022〕49号)。
关键词
粒子群算法
K均值聚类
柯西变异
遗传算法
马尔科夫生存状态
动态化
particle swarm algorithm
K-means clustering
cauchy variation
genetic algorithm
markovian survival state
dynamic
