改进协同粒子群算法的汽车涂装线调度

Improved Cooperative Particle Swarm Algorithm for Automobile Painting Line Scheduling

针对汽车涂装线的复杂调度问题,提出了基于改进协同粒子群算法的新调度方法,建立了汽车涂装线的最优调度模型。将协同粒子群算法生成的各子种群看成一个个单独的Agent,构成一个多智能体联盟,子种群Agent能感知和学习联盟中其他Agent的最优位置和自身信息,进行粒子的惯性权重自适应调整。通过相互竞争,协同进化,得到联盟最优值,解码获取最优的调度策略。最后通过实例仿真,验证了该算法在涂装线调度优化中的可行性和有效性。

To solve the complex scheduling problem in automobile painting line, a new method based on an improved cooperative parti- cle swarm algorithm was proposed. Firstly, the model of optimal scheduling was established. Then the sub - swarm produced by the co- operative particle swarm algorithm was treated as an agent, and a multi - agent alliance was composed by all of these agents. Mean- while, the inertia weight of particle could be adjusted adaptively by perceiving and learning the optimal position informations from them- selves and the other agents. Through competing with each other, agents could evolute cooperatively to find out the optimal value of alli- ance, and decode to get the optimal scheduling policy. Finally, the algorithm was simulated to prove its feasibility and effectiveness when be used in automobile painting line scheduling.