基于PSO-GRNN算法的物流运输风险预测

Prediction of Logistics Transportation Risks Based on PSO-GRNN

针对城市道路物流运输环节经常发生的车辆碰撞和货物爆炸等事故问题,从货运司机的个体感知视角出发,将“人-车-环境”三方面的风险因素作为输入,以广义回归神经网络(GRNN)为基础构建物流风险预测模型,并采用粒子群算法(PSO)和训练集对预测模型的光滑因子参数进行优化。经测试集验证,与BP神经网络模型和GRNN模型相比,PSO-GRNN预测模型的准确度提高了7.7%。结果表明:在训练集较少的情况下,PSO-GRNN预测模型也能达到预测准确度更高、稳定性更强的效果。

In this paper,to deal with the common occurrences in urban road logistics transportation links such as vehicle collisions and cargo explosions,etc.,based on the individual perception of the truck driver and with people,vehicle and environment as input factors,we built a logistics risk prediction model based on the generalized regression neural network(GRNN),and optimized the smooth factor parameters of the prediction model using the particle swarm optimization algorithm(PSO)and the training set.Through verification,it was shown that,compared with the BP neural network model and the GRNN model,the PSO-GRNN prediction model was 7.7%more accurate.The result also showed that the PSO-GRNN prediction model could achieve higher prediction accuracy and more stable effect despite smaller training set.