应用于铣削参数优化的粒子群和遗传交互算法  

Mutual studying algorithm integrated PSO and GA and its application to optimal parameters for milling computer engineering and applications

作　　者：马超[1] 蔡军[1] 杨飞[1] 崔彬[1] 

机构地区：[1]国电南瑞科技股份有限公司,南京210061

出　　处：《计算机工程与应用》2015年第16期252-258,共7页Computer Engineering and Applications

摘　　要：针对工程领域中的非线性、多极值和多维度等复杂优化问题,提出把遗传算子引入粒子群算法中,采用粒子搜索变异,交互学习的方法。方法综合了粒子群算法原理简单、搜索速度快,遗传算法全局搜索能力强的特点,实现了算法避免陷入局部最优解,以获得较高的精度和执行力。通过对比分析,此交互学习策略在求解精度、效率和处理多种复杂度问题等方面都有优越性,特别适用于精确求解和解决复杂优化问题。实例证明,算法可以解决基于机械动力学理论的铣削参数优化中非线性、多极值、多维度的工程问题。There are many non-linear,multi-extremum and multidimensional complicated problems in the applications of the engineering field.This paper puts forward the Genetic Algorithm（GA）into the Particle Swarm Algorithm（PSO）,and uses the method of mutual learning to solve those problems.This method integrates the particle swarm algorithm＇s simple theory and quick convergence with the genetic algorithm＇s global search ability to get higher convergence precision,stronger execution and avoid falling into local optimal solution.The comparative analysis results show the parallel learning strategy has great advantages in terms of accuracy,efficiency and processing ability of different complexity problems.This algorithm is especially applicable to solve accurately and complex problems.Example shows that this algorithm can solve the nonlinear,multiple maximum and multi-dimension engineering problem existed in the milling parameters optimization solved by the machine dynamics theory.

关 键 词：粒子群算法 遗传算法 交互学习 机械动力学 铣削参数优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP301.6[自动化与计算机技术—控制科学与工程]