Abstract: According to characteristics of process production in steel plant, a constraint satisfaction model for short-term production capacity planning was established based on constraints of finite production capacity, inventory and technology process. As the Constraint Programming(CP)may fall into an endless backtracking when the problem's scale was large and difficult to find feasible solutions, a constraint-based network production capacity calculation method was proposed. The pretreatment on available capacity of order/product's production network was obtained by using this method, and this capacity was considered as a constraint aided CP to cut the searching space, thus the solution was found in short time to do ability commitment and production scheduling of order/product. The experimential results showed that the complexity of algorithm was in a linear growth with the increasing of problem's scale, and the algorithm could be generally applied in various modes of production.

Key words: iron and steel plants, short term production capacity planning, constraint programming, network capacity, algorithms

摘要： 根据钢铁企业流程式生产的特点，以有限生产能力、库存和工艺制程为约束，利用约束规划方法建立了短期生产能力计划的约束满足模型。针对常规约束规划求解大规模问题过程中不存在可行解时大量回溯对求解效率产生的影响，提出一种基于约束的网络生产能力计算方法，首先对产品/订单的生产流程网络进行预处理，快速计算可用能力，然后利用网络生产能力约束辅助约束规划裁剪搜索空间，实现在短时间内找到解并进行订单/产品的能力承诺以及生产排程。通过算例实验证明了算法的复杂性与问题规模增长呈近似线性关系，而且可适用于不同生产模式，算法具有通用性。

关键词: 钢铁企业, 短期生产能力计划, 约束规划, 网络能力, 算法