Research On Short-term Traffic Flow Prediction Model Of Wavelet Neural Network Optimized By Evolutionary Algorithm And Error Compensation

In intelligent transportation system,short-term traffic flow prediction has always been the focus of domestic and foreign scholars.In order to improve the prediction accuracy,more and more combination models are applied in the field of short-time traffic flow prediction,in which the wavelet neural network model combines the advantages of wavelet analysis and neural network,and it has a good effect on the prediction of short-time traffic flow.However,the initial weights and parameters of the wavelet neural network are set randomly,and gradient descent algorithm is adopted in the learning and optimization process.The optimization is easy to fall into local extreme values,resulting in low prediction accuracy and insufficient prediction stability of the model.Therefore,in this paper,Genetic Algorithm(GA),Method Evolutionary Algorithm(MEA)and Error Compensation Method(ECerror)are adopted to optimize the wavelet neural network and build a model for short-term traffic flow prediction.The main research work of this paper is as follows.(1)Short-term traffic flow data preprocessing.In this paper,short-term traffic flow data of an observation station in Beijing were selected for data restoration and phase space reconstruction of the data set.Finally,the data were normalized to ensure that the obtained data could be used for model input.(2)Genetic algorithm and improved genetic algorithm optimize weights and parameters.In this paper,Wavelet neural network model(WNN)is constructed.The prediction accuracy and the prediction stability need to be enhanced.Therefore,this paper not only introduces Genetic Algorithm to construct the Wavelet neural network model based on Genetic Algorithm(GA-WNN),but also uses fuzzy clustering search method to optimize the Genetic Algorithm,and constructs the Improved Genetic Algorithm prediction model based on Improved Genetic Algorithm(IGA-WNN),and conducts simulation experiments.The results show that the improved model improves the accuracy of short-term traffic flow prediction.(3)The combination of evolutionary thinking algorithm and error compensation method.In view of the gradient descent Algorithm of weights and Wavelet parameter is sensitive to the initial value,it is easy to fall into extreme value problem of faults,this paper uses the mind Evolutionary Algorithm to optimize the Wavelet neural network,and the introduction of Error Compensation Method,build Wavelet neural network prediction model based on mind Evolutionary Algorithm and Error Compensation(MEA-EC-WNN).Simulation results show that compared with the results of GA-WNN prediction model and IGA-WNN prediction model,the accuracy of MEA-EC-WNN prediction model is higher,and its ECerror reaches 0.967.