Neural Network to Solve the Static Wagon-Flow Allocation

In marshalling station of railway, the problem of static wagon-flow allocation(SWA) can be transformed into the problem of balanced transportation of fixed cost by constructing the network model of SWA. After analyzed a few network model distinguish methods, the neural network has been trained is given. The algorithm gives the initial value to virtual arriving trains firstly, apply the learning rules of neural network to guarantee the outbound trains full loaded in the process of calculating, and finally work out the maximum virtual outbound trains. The example demonstrates that this algorithm can solve the SWA problem in a large scale within a reasonable time which provides a new approach to the SWA problem.