摘要
研究铁路客运量的优化管理,可以为国家资源分配提供依据,铁路客运量预测对铁路企业的经营决策也有着良好的指导意义,针对传统RBF神经网络极易陷入局部最优问题,为了提高铁路客运量的预测精度,提出一种基于遗传优化RBF神经网络的铁路客运量预测方法(GA-RBFNN)。GA-RBFNN首先用遗传算法优化神经网络的参数,并在遗传进化过程中保留最优个体的方法,选择参数的最优解来建立最优预测模型。以我国1985-2008年铁路客运量数据对GA-RBFNN进行仿真,结果表明,采用经遗传算法优化后的RBF神经网络模型比传统RBF神经网络有更高的预测精度和收敛速度,适用于铁路客运量等非线性预测问题,具有较高的预测精度和应用价值。
In order to improve the forecasting performance of neural network and forecast the railway passenger capacity rate more accurately,a neural network optimized by Genetic Algorithm approach was proposed for forecasting railway passengers.capacity accurately.Aimed at the problem that BP algorithm is usually trapped to a local optimum and has a low speed of convergence,the Genetic Algorithm is used to optimize the connection weights of neural network.In the evalution processes the best individual is reserved and selected to build the forecasting model.The data from 1980 to 2008 is used to testify and analyze the performance of the proposed model.The result shows that the proposedmethod can obtain amore accurate result than the traditional RBF neural network.Therefore the RBF neural network improved by GA is suitable to solve nonlinear problems such as prediction of railway passenger capacity,and has high accuracy and application value.
出处
《计算机仿真》
CSCD
北大核心
2010年第10期168-170,174,共4页
Computer Simulation
关键词
铁路客运量
神经网络
遗传算法
预测
Railway passenger capacity
Neural network
Genetic algorithm
Forecasting
