基于遗传搜索算法热轧合同编制优化仿真研究

Research on Optimization and Simulation of Planning Hot Rolling Orders by Genetic Search Algorithm

为解决钢铁企业多品种、小批量的热轧合同编制优化问题,针对规模大、约束复杂难以建模及求解等难点,以半旬为基本时间单位,在考虑各钢种炼钢能力、轧制能力等约束条件的基础上,建立以合同的提前期、拖期惩罚最小,各工序产能利用均衡,相邻排产合同的工艺约束惩罚费用最小以及各半旬的炼钢余材最少为优化目标的0-1非线性整数规划模型。由于所建模型具有多旅行商问题结构的特征及模型中约束条件复杂、数据规模较大,采用分段整数编码和启发式修复策略的遗传搜索算法进行求解。通过对实际生产数据进行仿真,验证了所提模型和算法的有效性,为科学合理地编制热轧合同计划提供了有效的解决方法。

To solve problem of planning hot rolling orders with the characteristics of multi-varieties and small-batch in steel enterprises,in view of the difficulty in modeling and making solution due to large scale and complex constraints,a multi objective nonlinear integer model was established based on consideration of steel making capacity and hot rolling capacity constraints.The objects of the model are to minimize the earliness/tardiness penalty cost,process constraint penalty costs between adjacent orders and surplus materials in steel-making process and to balance utility of capacities.An improved genetic-variable neighborhood search algorithm with sub-integer encoding method and heuristic repair strategy was proposed.Finally,the effectiveness of the proposed model and algorithm was verified by simulation based on actual production data.The proposed method can produce a more scientific way for hot rolling order planning.