基于神经网络的短时交通流量预测分析-交通运输规划与管理专业论文.docx

II II Abstract Abstract: Intelligent Transportation Systems (ITS) is an effective and a new way to ease urban traffic pressure and reduces environmental pollution. Traffic flow forecasting is an important part of ITS, and the short-term traffic forecasting is the main study content of the traffic flow forecasting. It can ensure the smooth flow of the intersection by accepting the short-term traffic forecast information timely. Therefore, how to obtain the accurate short-term traffic forecast information is the key to ensure the effective operation of the traffic. According to the highly nonlinear and uncertain characteristics of urban road traffic and the methods of the previous traffic flow forecasting, this paper establishes the short-term traffic flow prediction model based on the research of BP neural network, training the BP neural network by the LM learning algorithm, and then forecasts the city intersection traffic flow, the simulation example demonstrates the feasibility of the algorithm. In order to overcome the shortcomings of the BP neural network that is easy to fall into local minimum points caused by the improper selection of the initial parameters, this paper optimizes the initial parameters of the BP neural network by the intelligent optimization algorithm, and introduces the basic principles of the particle swarm optimization and genetic algorithm and their improved algorithms, then optimizes the initial weights and thresholds of BP neural network by the intelligent algorithms, and establishes the short-term traffic flow prediction model based on PSO-BP, MPSO-BP, IAGA-BP neural network forecasting model on this basis. After the intelligent algorithms optimize the neural network, the simulation shows that the prediction accuracy is improved. Finally this paper makes a summary of the prediction algorithm, and introduces the idea of genetic algorithm to the particle swarm optimization, proposes the APSO algorithm which combines the advantages of the PS