针对生产调度中的多目标混流装配线排序问题,建立以最小化超载时间、产品变化率与总切换时间为优化目标的数学模型,并提出一种改进的多目标粒子群算法求解。该算法采用基于工件的编码方式,并提出新的解码方法;应用Pareto排序和小生境数评价个体,在此基础上形成了一种新的适应度函数。在个体最优解的更新中,为避免最优解丢失,对非支配粒子与支配粒子采用差异化方法更新。此外,运用两种策略解决粒子群算法过早收敛的问题:在个体最优解的更新中引入模拟退火思想,并将全局最优解的选择扩大到整个种群。通过数值算例研究了算法的收敛性、分布性和执行效率,结果表明了所提算法的优越性。
Aiming at the multi-objective sequencing problem in mixed model assembly lines,a mathematical model was proposed with the optimization objectives of minimizing total utility work,total production rate variation and total setup cost.Besides,an improved Multi-Objective Particle Swarm Optimization(MOPSO) was proposed to solve the model.In the algorithm,job-based coding was introduced and a new decoding method was put forward.Pareto ranking and niche count were employed to evaluate an individual,and a new fitness function was formed on these basis.In the update process of personal best,non-dominated particle and dominated particle were differentially updated so as to avoid lose of optimal solution.Furthermore,two strategies were adopted to overcome the drawback of premature convergence in particle swarm optimization: Simulated Annealing(SA) was introduced into the update of personal best and the selection of global best was extended to the whole swarm.Several numerical examples were presented to study the convergence,distribution and efficiency of the proposed algorithm,and the results showed the superiority of the algorithm.