中文摘要：

针对冷轧生产流程多阶段库存和批次多流向的特点,在满足工艺约束的基础上建立了以合同超期和提前的总惩罚费用最小、中间库存量超出和不足的总惩罚费用最小并确保成品库成本最低为目标的多阶段多流向冷轧生产计划模型,采用改进的遗传算法进行求解。算法通过参数的合理设定,解决了交叉、变异概率在一定区间内取固定值、断点处函数值跳跃大和出现不合理值的问题,使两种概率可根据适应值进行自适应调整。以国内某冷轧企业为案例对模型和算法进行了验证,实验结果表明该冷轧生产计划模型能够保证合同交货期和降低库存成本。

英文摘要：

Aiming at the characteristics of multi-stage inventory and multi-flow directions in cold rolling production process, based on meeting the process constraints, a multi-stage inventory and multi-flow directions production planning model was established, which took undue or overdue contracts, sufficient or insufficient intermediate inven- tory and minimum inventory costs of finished products as the objectives. The model was solved by an improved ge- netic algorithm. Through setting reasonable parameters, the occurrences where crossover probability and mutation probability took fixed values within a certain range and the function values appeared big jumps at some breakpoints or unreasonable value were avoided. These two probabilities could be adjusted adaptively by the fitness of individu- als. A case study was presented to verify the effectiveness of the model and algorithm, and the result showed that the presented model could guarantee due dates of contracts and reduce inventory costs.