摘要
根据复杂产品装配规划问题的特点和要求,提出了一种求解装配序列规划(assembly sequenceplanning,ASP)问题的粒子群优化算法,将通常用于连续空间优化的粒子群算法成功扩展到ASP领域.算法根据ASP问题决策解的特点,在排序空间定义了微粒的位置和速度以及相关的各种操作.针对基本粒子群算法容易陷入局部最优的缺点,采用新的学习机制,增强了算法的寻优能力.基于干涉矩阵、连接矩阵和支撑矩阵建立了以装配可行性、装配体稳定性和装配方向改变为评价指标的目标函数.最后通过实例分析验证了该算法的有效性.
According to the characteristics and demands of assembly sequence planning (ASP) of complex products, the particle swarm optimization (PSO) algorithm, which is used mainly to optimize the spatial continuity, is extended to solve the ASP problem. The algorithm redefines the particle's position, velocity and relevant operations in sequencing space in accordance to the characteristics of solution. To rise above the deficiency that PSO algorithm is easy to fall into local optimization, a new learning mechanism is taken up to improve the optimizability of the algorithm. Based on the interference matrixes, connection matrix and support matrix, the geometrical feasibility, assembly stability and the occurrence of changing the assembly direction are all taken into account as the evaluation indices to form an objective function. The validity and feasibility of the proposed algorithm have been verified via exemplification.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2010年第2期261-264,共4页
Journal of Northeastern University(Natural Science)
基金
解放军总装备部武器装备预研基金资助项目(9140A18010207LN0101)
关键词
装配序列规划
智能优化算法
粒子群优化
组合优化
ASP(assembly sequence planning)
intelligent optimization algorithm
PSO(particle swarm optimization)
combination optimization