摘要
针对传统Job-Shop数学模型忽略返工及重加工的因素,构建了考虑该情形下的Job-Shop调度数学模型及相应的求解算法。该模型详细分析了返工及重加工的流程,对问题的定义做了进一步推导,模型以总加权拖期最小为目标,并提出一种改进的遗传算法对该模型进行求解。针对该调度情形,对算法中染色体的编码、种群初始化进行改进。种群数据的仿真实验表明,与传统遗传算法相比,改进后的算法在收敛速度、求出的最小总加权拖期方面均优于前者。最后通过对10×10实例调度方案求解及仿真,并与作业车间实际调度结果比较,模型仿真所得总加权拖期小于实际计划调度结果的46%,本模型得出的调度方案是实用且有效的。
A Job-Shop scheduling is addressed with the consideration of stochastic rework and reprocessing while the traditional Job-Shop mathematical model ignored it. The objective of the model is to minimize the Total Weighted Tardiness （TWT） in these job shops. To solve the problem, a modified genetic algorithm is proposed with coding and population initialization improved. Five groups of population data for simulation experiments show that compared with the traditional genetic algorithm, the improved algorithm is better in convergence speed and the target value of TWT. Finally, through simulating the instances of scheduling scheme, and comparing with the actual Job-Shop scheduling results, the TWT of model simulation is less than 46% of the actual project scheduling results, and the scheduling scheme model proves practical and effective.
出处
《工业工程》
2015年第5期127-133,147,共8页
Industrial Engineering Journal
基金
广东省教育部产学研结合资助项目(2011 B090400160)
