带软时间窗的联盟运输调度问题研究

为解决允许使用不同类型车辆和多层次交通网络的带软时间窗的联盟运输调度问题，在建立数学模型的基础上，利用改进的蚁群算法求解。首先，为了克服蚁群算法最优解不稳定和易陷入局部最优等缺点，按经验将选择策略分为3个阶段，每个阶段选用相应的转移概率，并根据信息素浓度与挥发速度的关系自适应调整信息素挥发因子；其次，为了解决蚁群算法不易发现可行解的问题，从构造3类分支回路和处理遗漏客户点人手，构造了联盟运输调度问题的可行解。仿真计算表明，该算法简明有效。

Ant colony algorithm for allied vehicle routing problems with soft time windows

To deal with Allied Vehicle Routing Problems(AVRP) of different vehicles and multi-hierarchy traffic networks,based on its mathematical model,the improved Ant Colony Algorithm(ACA) was presented.Firstly,to overcome the weakness that ACA was easily trapped into local optimal solution and unstable optimal solution, the selection strategy was divided into three phases and transfer probability was chosen for each phase accordingly,pheromone volatilization gene was also self-adaptively adjusted according to the relationship between strength and volatilization velocity of pheromone.Secondly,to overcome the weakness that ACA was difficult to find feasible solution,some strategies of establishing three types of branch circuits and processing skipped clients were presented to construct the feasible solution of AVRP.Finally,simulation computation showed its effectiveness.