居民出行产生量BP神经网络预测方法

居民出行产生量预测是交通需求分析的重要内容之一,预测结果是确定各类城市交通设施发展规模及布局规划的重要依据.通过分析人工神经网络的作用机理和居民出行产生量的影响因素,建立了居民出行产生量预测的四层BP神经网络模型,以土地利用作为输入神经元,以交通区居民出行产生量作为输出单元,以赣州市城市综合交通规划交通调查数据对模型进行了标定与检验,并与出行次数法和回归分析法进行了比较,结果表明BP神经网络模型具有较高的预测精度.

A forecast method for trip production based on BP neural network

Trip production forecast is one of key components of traffic demand analysis,which directly determines the scale and layout of different urban traffic facilities.The mechanism of artificial neural network (ANN) and influential factors of trip production were analyzed.therefore,a four-layer back-propagation neural network (BP neural network) model was set up to forecast the trip production,in which the input neurons are land-uses of different traffic zones and the output is trip production.Meanwhile,the model was calibrated and testified with traffic survey data from the urban integrative transportation planning of Ganzhou city.Furthermore,the results were compared with those obtained from trip production rate method and multiple linear regression method,It is showed that forecast precision of the BP neural network is relatively high.