基于RBF神经网络的不利天气道路通行能力计算

不利天气下影响城市道路通行能力的各种因素都具有随机的、非线性,采用常态条件下修正理论通行能力的计算方法是不适合的.文章结合RBF神经网络模型方法能够良好地分析出随机的、非线性的特点,对路网组成单元进行重新划分,选定不利天气下道路通行能力的影响因素,建立了道路通行能力计算的RBF神经网络模型.并依据哈尔滨市暴雨天气下道路的实际情况进行了算例分析,计算的道路通行能力与实测数据最大误差为-1.16%,验证了模型的可行性和有效性.

Road Capacity Calculation under Adverse Weather Based on RBF Neural Network

Various influencing factors of urban road capacity under adverse weather are stochastic and nonlinear. Therefore, it is unsuitable to adopt normal amendment method to calculate the road capacity. Based on the RBF Neural Network, which can analyze the stochastic and nonlinear characteristics, the components of road network were redivided and the influencing factors were re-selected. Then, the RBF Neural Network model was built for calculation of road capacity. An example was provided based on the actual situation of Harbin City under rainstorm. The maximum error between the calculation results and the observation data was -1.16%. Thus, the feasibility and validity of the model were validated.