冷轧多机组合同优化排产模型及算法(上)

来源期刊：冶金自动化2010年第4期

论文作者：王志刚 邬静东 赵珺 王伟

文章页码：10 - 15

关键词：冷轧； 生产调度； 时间窗推理； 旅行商问题； 蚁群算法

Key words：cold rolling； production scheduling； time-window reasoning； traveling salesman problem； ant colony optimization

摘    要：为解决钢铁冷轧多机组的排产问题,建立了一个实现合同生产计划和作业调度的整体优化模型。模型以最小化各机组合同延期惩罚和生产类型切换虚拟成本为优化目标。首先,针对上下游机组的物流平衡及保证交货期问题,根据合同钢卷在生产流程各机组间的传递构造基于时间窗的推理算法,确定钢卷的合理加工范围。在传递时间窗下各机组生产调度可以不必考虑机组的前驱约束。然后,根据合同交货期、产品种类以及工艺约束等将合同钢卷在各机组划分为不同的粗计划类型。同一类型粗计划合同钢卷的批次作业调度可以归结为多约束无返回多旅行商问题,通过一种启发式分组蚁群算法优化求解。最后,将机组的粗计划合同调度抽象为带时间窗旅行商问题,采用一种改进的自适应蚁群算法求解。通过现场实际数据试运行,结果表明所建模型与算法是有效可行的,为冷轧企业多机组的合同计划和调度提供了合理指导。

Abstract: To solve the problem of order planning and coil scheduling for a multi-process cold rolling line in a steel enterprise,an optimization model is proposed.Objective of the model is to minimize order delay punishment and virtual switch cost in each process.Firstly,in order to address the problem of line balancing and due date conflict,a time-window based reasoning algorithm is used according to transfer of coils between production processes.Front-drive constraints can be neglected for scheduling of each process under transfer time windows.Then,coils are classified as different campaign types at each process according to their due dates,product type and technical constraints.Batch coil scheduling of each campaign type is mapped as CNRMTSP(Constrained No-Return Multiple Traveling Salesman Problem),and an heuristic team ant colony system is designed to optimize its solution.Finally,scheduling of campaigns can be Abstracted as TSPTW(Traveling Salesman Problem with Time Windows).An improved adaptive ant colony algorithm is developed to solve the problem.The running results with field data show that the proposed model and algorithms are feasible and effective,and can provide reasonable guidance for production planning and scheduling in the cold rolling plant.