基于改进粒子群算法的路径优化问题研究  

Research on Path Planning Problems Based on Improved Particle Swarm Optimizer

作　　者：梁静[1] 宋慧[1] 瞿博阳[2] 毛晓波[1] 

机构地区：[1]郑州大学电气工程学院,河南郑州450001 [2]中原工学院电子信息学院,河南郑州450007

出　　处：《郑州大学学报（工学版）》2014年第1期34-38,共5页Journal of Zhengzhou University（Engineering Science）

基　　金：国家自然科学基金资助项目(211000531605);中国博士后科学基金特别资助项目(2012T50639);教育部高等学校博士学科点专项科研基金(20114101110005);河南省科技攻关项目(132102210521)

摘　　要：把动态约束机制及改进的粒子群优化算法(含交叉策略的动态多组群粒子群优化算法)与Bezier曲线结合来实现路径优化.用三次Bezier曲线来描述路径,测试不同算法下得到的曲线在路径优化问题中的特性,通过比较得出不同算法的优劣性.实验结果表明:动态多组群粒子群优化算法克服了标准粒子群优化算法易于早熟和陷入局部最优的特点,动态约束提高了交叉策略下算法的寻优性能,能够灵活地处理约束条件,克服了以往静态惩罚函数的缺点.This paper describes that dynamic constraint mechanism and improved Particle Swarm Optimizer （Dynamic MultiSwarm Particle Swarm Optimizer with Crossover） are combined to solve path planning prob lems combining with Bezier curves. Different algorithms are used to test characteristics of the curves in the path planning process and Bezier curve is used to describe the path. The results show that the DMSPSO over comes the phenomenon of precocious and easy convergence compared with standard PSO, and constraint mech anism improves performance of the optimization algorithm which is combined with crossover, can deal with the constraints flexibly and overcomes the shortcomings of static penalty function.

关 键 词：动态约束机制 动态多组群粒子群优化算法 BEZIER曲线 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]