基于约束满足的热轧批量计划模型与算法

将热轧批量计划问题作为一个约束满足问题处理，建立不确定计划数的VRPSTw约束满足模型．在求解过程中．先用约束满足的一致性技术过滤变量的值域，收缩搜索空间；然后用变量选择和值选择构造轧制计划的解．为变量赋值之后，实施约束传播，保证每块板坯只被访问一次并动态禁止子回路．在已有的解的基础上，应用基于禁忌的k-opt互换改进解的质量．数据实验证明模型和算法是有效的．

Model and algorithm for hot-rolling batch plan based on constraint satisfaction

The hot-rolling batch planning problem is treated as a constraint satisfaction problem. A VRPSTW constraint satisfaction model with uncertain rolling plan number is established. During the solving procedure, domains of variables are filtered by applying constraint satisfaction consistency techniques, and the solution is constructed by the variable selection and value selection heuristics. After the instantiation of a variable, constraint propagation is implemented, the constraint that each slab is visited only once is guaranteed and the subtour is dynamically forbidden. The quality of the solution is further improved by using tabu-based k-opt exchange. The numerical experiments show the validity of the proposed model and algorithm.