基于JIT的并行多机问题的病毒进化遗传算法

为了解决提高顾客对完工时间满意度的提前/拖期调度问题,提出了一种将病毒进化遗传算法和单机问题的有效算法相结合的求解方法。在该方法中,病毒进化遗传算法采用整数编码,简化了算法的实现难度,提高了算法的执行效率。每个染色体都代表完成零件加工的机器编码的组合,通过病毒种群对主种群的病毒感染,提高了逃脱局部极点的搜索能力。在每一代种群中,调用单机问题的有效算法,使达到用户满意度的零件数最大。数值仿真实验表明,该算法具有收敛速度快、优化效果好等特点,且适合于求解大规模问题。

Virus evolutionary genetic algorithm for solving a multi-objective job-scheduling problem on parallel machines based on JIT

In order to solve the earliness-tardiness job scheduling problem of improving customers' satisfaction toward completion time, an algorithm combining Virus Evolutionary Genetic Algorithm (VEGA) and single machine problem was presented. In this solution, due to VEGA's adoption of integral coding, it simplified the algorithm's accomplishing difficulty and enhanced the algorithm's execution efficiency. Each chromosome represented a combination of the machine coding about finished accessory processing, through virus generation's infection on the main generation, the searching ability of escaping from the local apices was increased. With each generation, the effective algorithm used for transferring single machine problem has lessened the amount of user satisfactory degree accessory to the maximal scale. Numeric examples demonstrated that the genetic algorithm carrying the property of fast convergence. The solution gained by the genetic algorithm was pretty well.