约束规划方法在原油混输调度中的应用

Application of constraint programming method in crude scheduling

应用约束规划方法建立炼厂原油混输模型,并通过搜索求解获得可操作的调度方案。约束规划允许以逻辑约束表达混输过程中的操作规则,使模型简洁直观,而对整型变量及非线性约束不敏感,则降低了模型的求解难度。选取了一个时间跨度为300小时实际调度问题作为案例,模型中包含约8000个变量和14000条约束,在普通PC机上计算求解,能够在30秒内获得调度方案,方案具有可操作性且优于人工方案,表明以约束规划方法解决原油混输调度问题是可行的。

Constraint programming （CP） method has been applied to construct crude feeding model for refineries. Feasible scheduling scheme can be acquired by solving the model with search strategies. In order to achieve a compact model, logical constraints enabled by CP are used to express the operation rules in crude feeding. CP models are insensitive to integer variables and nonlinear constraints, which reduces the difficulty of the solving procedure. An industrial scheduling problem with the horizon of 300 hours is introduced. Model for the problem is formulated with about 8000 variables and 14000 constraints. The model is solved on a common personal computer and a scheduling scheme is obtained in less than 30 seconds. The scheme is examined to be operable and proved to be more optimal than manual scheme, which demonstrates the feasibility of solving crude feeding scheduling problem with CP method.