摘要
提出了一种新的多目标组合运输物流量预测建模算法.以时间、领域、影响以及组合运输为基准,运用系统工程理论思想设计出一种四维的物流量影响因素模型,并运用结构方程模型对所建模型做了优化,提取出组合运输物流量的核心影响因素.在改进的神经网络算法的基础上结合遗传算法,提出了一种结合遗传算法的改进的神经网络新算法,弥补了改进的神经网络算法上的缺陷,在多目标组合运输物流量预测的实例应用中,该算法不仅有很高的预测精度,而且具有收敛速度快、运行稳定的特点.
A new method was brought forward for the modeling of multi-objective prediction on logistics volume of combined transportation. Based on the standard of time, field, influence and combined transportation, using systems engineering antilogy, a model of four-dimensional factors of logistics volume was designed and optimized by using structural equation model. The fatal influencing factors of logistics volume of combined transportation were distilled. A new advanced neural network arithmetic integrated with genetic algorithm was put forward to make up the limitation of advanced neural network, and applied in a example of multi-objective prediction on logistics volume of combined transportation. Results show that this advanced algorithm performs steadily with high precision and convergence speed.
出处
《北京航空航天大学学报》
EI
CAS
CSCD
北大核心
2006年第10期1209-1214,共6页
Journal of Beijing University of Aeronautics and Astronautics
关键词
多目标预测
组合运输物流量
结构方程模型
遗传算法
神经网络
multi-objective prediction
logistics volume of combined transportation
structural equation model
genetic algorithm
neural network
