Abstract: In cargo circulation process, the current epidemic of vehicle scheduling is based on simple neural network model design, and thus make transportation cost waste. To this end, this paper proposes a non-full loads vehicle route optimization mining model based on improved neural network to solve the problem of non-full loads vehicle scheduling optimization in the process of transportation. Improved model by weighting the non-full loads vehicle length of time domain and the airspace of probability, constraining the neural network on steady state, building the non-full loads vehicle starting point and end point function equation improved algorithm distribution model, weighting time window of new model, to generate the distribution model of time window weighting and based on the new model, merges the improved neural network non-full loads mining vehicle model. The simulation results show that the mining model compared with the traditional calculation method of the neural network, can improve the profit of the non-full loads vehicle logistics, the effect is remarkable.

Key words: neural network, non-full loads vehicle, route optimization

摘要： 货物流通过程中，目前流行的车辆调度方式——基于简单的神经网络模型设计，造成运输成本的浪费。提出了一种基于改进神经网络的非满载车辆路线优化挖掘模型，来解决运输过程中的非满载车辆调度优化问题。改进的模型通过对非满载车辆时域长度和空域概率的加权、对神经网络稳定状态进行约束、建立非满载车辆起点和终点函数方程生成改进算法配送模型，并通过对新模型进行时间窗加权，合成了改进神经网络非满载车辆挖掘模式。仿真结果表明，该挖掘模型与传统的神经网络计算方法相比，能够提高非满载车辆路线选择效率和正确性，取得了较好的效果。

关键词: 神经网络, 非满载车辆, 线路优化