针对带有非正规性能指标的作业车间提前/拖期调度问题,基于工序编码方式,设计了一种适用于粒子、染色体等个体的三阶段解码方法。通过半主动解码操作,以增加初始调度解的搜索范围;采用插入式贪婪解码这类主动调度方法,以减少拖期零件子集的拖期惩罚成本。运用逆向主动解码操作来延迟所有零件的释放期,并在保证拖期零件完工时间不变的前提下进一步压缩提前惩罚成本。这种三层递阶解码方法有助于大幅度压缩提前/拖期惩罚总成本。将该解码方法应用于粒子群算法中的粒子解码过程,并通过粒子位置更新操作来实现调度解的迭代优化。通过120个随机调度算例的仿真实验,证明了该方法的有效性。
Aiming at the Job Shop earliness and tardiness scheduling problem with informal performance index,a new decoding method with three hierarchical structure applied to particle and chromosome was designed based on procedure code pattern.A semi-active decoding operation was employed to extend the search band of initial scheduling solution.The active scheduling method of insert greedy decoding was used for the tardy jobs to reduce the tardiness cost of jobs.The active decoding operator with reversal mechanism was operated to delay the due date of all parts,and the premature punishment cost was reduced further on the premise of tardy jobs' completion time unchanged.Thus,the total tardiness and earliness costs could be cut down owing to these three successive decoding procedures.This method was applied to particle decoding in Particle Swarm Optimization(PSO),and the iterative optimization of scheduling solution was realized by updating the particle position.The effectiveness of the method was improved by simulation result of 120 random scheduling examples.