Abstract: An efficient modeling mechanism for the dynamic scheduling problem of steelmaking-continuous casting production under the machine failures is established based on the Dynamic Constraint Satisfaction Problem （DCSP）. The modeling mechanism describes variable factors of production easily and accurately. The impact on scheduling by machine failures is mapped to the dynamic changes of DCSP model in variables, domains and constraints. Consistency measurement of the schedules before and after rescheduling is presented, to meet the requirement of different charge for time variables and machine variables assignment. Disturbances from machine breakdown are divided into 3 levels according to their impact on scheduling strategy, and then the mapping relationship between the disturbances of machine breakdown and DCSP model of scheduling is built. Different solving strategies are established according to the degree of the machine failure, and the machine variables are reassigned for the operations in steelmaking. Based on the constraint propagation, and through adjusting the starting and processing time, the conflicts on time variables are cleared in the algorithm. The validity of the model and the algorithm is testified by the simulation experiments.

Key words: steelmaking-continuous casting, dynamic scheduling, dynamic constraint satisfaction, machine failures

摘要： 针对机器故障下的炼钢-连铸动态调度问题，基于动态约束满足技术开发了能够灵活反映各种动态因素的建模机制。从变量、值域和约束三个角度将生产过程中的机器故障的影响映射为约束满足模型的动态变化；提出了重调度前后调度方案在时间安排和机器指派上的一致性度量方法，以满足不同炉次对时间和机器一致性的不同要求；将机器故障扰动按影响程度分为3个层级，建立了故障扰动与约束满足调度模型间的映射关系。根据炼钢、精炼阶段的机器故障扰动程度，制定不同的求解策略，并为机器指派变量赋值；基于约束传播技术，通过调整开工时间和柔性加工时间分步消解时间约束冲突。仿真实验表明，提出的模型和算法是可行和有效的。

关键词: 炼钢-连铸, 动态调度, 动态约束满足, 机器故障