基于粒子群模拟退火算法的协同制造任务链构建  

Task Chain Construction of Collaborative Manufacturing Based on Particle Swarm Optimization and Simulated Annealing Algorithm

作　　者：陈君[1] 田锡天[1] 张振明[1] 耿俊浩[1] 董思洋[1] 

机构地区：[1]西北工业大学现代设计与集成制造技术教育部重点实验室,西安710072

出　　处：《计算机测量与控制》2013年第4期1005-1007,共3页Computer Measurement &Control

基　　金：国家863计划资助项目(2007AA040503);国防基础科研计划资助(A0520110036)

摘　　要：为了解决网络化制造中的任务分配与资源配置脱节的问题,从制造任务与制造资源相互协调的角度出发,建立了基于时间、成本和质量为多目标的协同制造任务链数学模型;考虑到粒子群算法后期由于粒子趋向同化使其容易陷入局部最优,引入模拟退火思想,设计了基于粒子群模拟退火的模型求解方法;该方法不仅汲取了粒子群算法快速收敛的优点,并通过模拟退火的降温过程来提高算法的进化速度和精度,保持较好的全局搜索能力;最后的仿真实验表明,所提出的方法收敛速度快,寻优能力强。To solve the disjunction problem between task assignment and resource allocation in networked manufacturing, a multi-ob-jective mathematical model of task chain was proposed in the view of coordination between tasks and resources, which set time, cost and qual-ity as targets. In consideration of stagnation phenomenon caused by assimilation tendency of particles in the later phase of particle swarm op-timization （PSO）, introducing the idea of simulated annealing （SA）, a collaborative heuristics based on PSO and SA was designed for this model. The provided algorithm not only possessed rapid convergence merit of PSO, but also kept strong global searching ability through im-proving evolution rate and precision by temperature decreasing procedures of SA. Finally, a simulation experiment is carried out by using the proposed algorithm, the results shows fast convergence and strong optimization ability.

关 键 词：网络化制造 任务分配 多目标 粒子群 模拟退火 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论] TP391[自动化与计算机技术—计算机科学与技术]