基于Hopfield神经网络的战时物资筹措模型

Raising of Wartime Material based on Hopfield Neural Network

战时物资筹措是国防动员的一个重要组成部分,它实质上是一种基于约束的最短路径问题,Hopfield神经网络曾被用来解决最短路径优化问题。采用连续型Hopfield神经网络求解战时物资筹措问题,在已知供需量、通行参数等数据的情况下可快速确定物资筹措方案。这种方法针对战时物资筹措的特殊性如时间严格、道路通行受阻等具体情况具有很好的适应性。实验结果证明在问题规模较大的情况下模型具有收效速度快、计算结果准确性高等优点。

Raising of wartime material is a key to national defense mobilization,which is essentially a Constraint-based Shortest Path problem.Hopfield neural network has been used to solve the shortest path optimization problem.In this paper,a Raising approach of wartime material using continuous Hopfield neural network is proposed,which can work out the scheme of Raising quickly with given supply and requirement capacity and other parameters.It is adaptive to Raising specialities such as strict time,blocked roads and so on.Experimental result shows that the speed of convergence is faster and the accuracy of calculated result is higher when problem is larger.