基于约束规划的一类排序问题通用求解方法

Refactoring model and generalized method for a type of scheduling problem based on constraint programming method

多约束排序问题是生产调度中常遇到的问题,传统的优化模型及方法在适应约束改变等方面存在诸多不足。鉴于此,将多约束排序问题定义为约束满足问题,系统设计时将模型定义与求解算法分离,利用约束规划平台的基本约束构建特定领域的抽象约束库,形成可重构的多约束排序问题通用求解框架。应用时,根据问题需求不同可利用抽象约束库快速重构优化模型,针对重构的优化模型配置相应的求解算法即可实现问题求解。应用结果表明,提出的方法通用性强,可满足实际应用的要求。

The multi-constrained scheduling problem is often encountered in production scheduling field, and it is difficult for traditional optimization models and methods to adapt the variable production constraints. Based on the constraint programming, the problem is defined as a constraint satisfaction problem. A refactoring model and a gen- eralized method are proposed to solve the problem. In the generalized solving framework, the model and the algo- rithm are designed separately. With the constraint programming platform, an abstract constraint library for the spe- cial domain is established. According to the different production environment, by linking the constraints library the refactoring model is quickly established. The optimization model is solved by the constraint programming approach and the appropriate search algorithm. Application results show that the proposed method is versatile and the results meet the requirements of practical application.