不确定条件下编组站动态配流模型及算法研究

Model and Algorithm of Dynamic Wagon-flow Allocating in a marshalling yard under Uncertaint Conditions

由于编组站解、编作业时间存在一定的波动性,将其作为模糊变量,用变量的λ悲观值表示在一定置信水平下的解、编作业时间,以阶段内出发车辆数最大为目标,建立不确定条件下的编组站动态配流模型。通过定义不确定条件下的可解集合、待解集合和选解集合将动态配流问题映射为方案树,从而设计一种基于蚂蚁系统的非确定性树搜索算法。由于改进了蚂蚁系统的选择策略和信息素更新,并在每次转移过程中对模型的约束条件进行判断,提高了解的性能和算法的收敛速度。算例表明,该算法能够较快地搜索到有利的全局方案。

Providing that the time of break up and formation at a marshalling station are set as fuzzy variables and expressed by λ,the pessimistic values of the variables,under certain confidence level and the maximum number of vehicles departed according to a stage plan are set as the target,the dynamic wagon-flow allocating model is built under uncertain conditions.The dynamic wagon-flow allocation is mapped into the scheme tree be defining the possible break up set,waiting set and choice set,so the approximate nondeterministic tree search algorithm based on the ant system is designed.Route selection improvement,pheromone updating and judgment of constraint conditions in each process of transfer enhance the efficiency of the solution and the convergence speed of the algorithm.The example demonstrates that the algorithm can find out a good global scheme in a relatively efficient way.