一种需求不确定的有容量网络设计问题求解方法

An Approach for Capacitated Network Design with Uncertain Demand

针对需求不确定的有容量网络设计问题,通过引入缺货成本(understocking cost)和过剩成本(overstock-ing cost),给出了一个两阶段的随机混合整数规划模型,并提出了基于Benders分解和分支定界新方法.该方法在解决电讯网络、物流运输网络中都有广泛应用背景.

It considers a network system with uncertain demand such as telecommunication network. In order to minimize the expected cost, understocking cost and overstocking cost is introduced and then a two-stage stochastic and mixed integer programming model is brought forward. Benders decomposition algorithm based on the diea of delayed constraint generation has been used here to solve the model. Branch and bound method is used when dealing with the integer variables. The approach is especially useful for the large-scale problem.