摘要
通过考虑带有并行机的Job Shop调度过程中设备故障对生产的影响,以最大完工时间和任务提前/拖期时间加权最小为目标建立数学模型。基于滚动时域优化基本框架,设计改进的事件和周期混合驱动的重调度机制。改进的事件驱动机制能过滤掉不必要的重调度,周期驱动机制能跟踪系统的变化,二者结合充分发挥两类驱动机制各自的优势。针对基本蚁群算法(Ant colony system algorithm,ACS)搜索时间长、易陷入局部最优解的缺点,引入精英蚂蚁策略和最大最小蚂蚁机制,设计改进蚁群系统算法(Improved ant colony system algorithm,IACS)。通过与ACS、遗传算法(Genetic algorithm,GA)的对比试验验证了IACS算法的优越性。该方法作为离散制造车间生产调度系统的组成部分在上海某军工企业生产车间实施,用实际生产数据验证动态调度方法的有效性。
Analyzed the impact of machine breakdowns in job shop schedule problem with parallel machines, a mathematical model with minimizing the weighted value of makespan and early due date as optimization goal is established. Based on the basic framework of the rolling time domain optimization, a hybrid driven schedule mechanism combining events and cycle is designed. Improved event driven mechanism filters out unnecessary scheduling, meanwhile, cycle driving mechanism can track system changes. The combination can give full play to their strengths. Due to the long searching time and easy falling into local optimum solution of Ant colony algorithm（ACS）, an improved ant colony system algorithm（IACS） mixing elite ant strategy and max-min ant mechanism is designed. The contrast experiment with ACS and genetic algorithm（GA） verified the superiority of IACS. The dynamic scheduling method is implemented as a constituent part of the scheduling system in discrete manufacturing workshop of a military enterprise in Shanghai, and the validity is verified by the production data.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2013年第14期182-190,共9页
Journal of Mechanical Engineering
基金
国家自然科学基金(60934008)
国家高技术研究发展计划(863计划
2012AA040907)资助项目
关键词
并行机
混合驱动
重调度
蚁群算法
Parallel machines Hybrid driven Dynamic schedule Ant colony algorithm
