基于改进遗传算法的铸造造型任务批调度模型

A Batch Moulding Scheduling Model in Foundry Based on Improved Genetic Algorithm

以单件小批量生产模式为主导的铸造生产具有订单种类多样、产品制造周期长、车间自动化程度低等现象,针对铸造企业客户订单多材质、铸件产品多类别以及造型熔炼多约束的特点,建立了一个以造型任务总完工时间最小的铸造造型任务批调度模型,并提出了一种改进的遗传算法对模型进行求解。算法设计了一种基于单件与砂箱类型的双层编码方案,在初始化阶段通过结合批首次匹配(BFF)规则进行分批,以提高初始种群的质量,在迭代阶段设计了一种基于批次交换的局部搜索方法,以避免算法陷入局部次优解。最后通过对某铸造企业的实际生产数据进行案例分析,验证了所提模型的有效性和算法的优越性。

The casting production,which is dominated by the production mode of single piece and small batch,has various orders,long production cycle and low degree of automation in workshop.According to the characteristics of multi alloys of orders,various casting items and multi constraints in foundry,a batch moulding scheduling model to minimize the makespan is proposed,and an improved genetic algorithm is presented to solve the model.A bi-layer coding scheme based on casting items and flask types is designed.In the initialization phase,and the batch-first-fit(BFF) heuristic method is applied to improve the quality of initial population;in the iterative phase,a local search approach based on batch exchange is introduced to help the algorithm escape from local optimum.Finally,a case study of a typical foundry enterprise is carried out to verify the effectiveness of our proposed model and the superiority of algorithm.