Improvement Of Two Kinds Of Optimization Algorithms And Their Application Research

With the rapid development of national economy and transportation infrastructure,transportation planning and guidance has become a hot issue of research in intelligent transportation system.For traffic planning and guidance,to solve the key problem is how to use the historical traffic data to predict the future traffic flow of each time period.The short-term traffic flow prediction at the core of intelligent transportation system(ITS)and the foundation,has increasingly become the focus of attention and research.According to the short-term traffic flow prediction problem.In this paper,a wavelet neural network(WNN)prediction model optimized by improved particle swarm optimization(IPSO)is proposed and analyzed.Stock market prediction problems related to financial time series forecasting is considered to be a challenging task.In this area there have been more involved in researches of the support vector machine(SVM).According to financial time series prediction problem,this paper adopts the improved fuzzy information granulation(IFIG)and SVM combining forecasting model,the related information into a financial time series,and then establish a reasonable particle to retain information,and then to make a prediction of financial time series.The main research work of this paper is as follows:1.Aiming at the limitations of PSO,such as the fact that PSO is prone to local extrema,and the improved from three aspects:the introduction of uniform random search operator,the adaptive inertia weight and the learning factor,and the performance of pso is tested on six benchmark functions.2.Traditional WNN forecast model,usually by using the method of one-way gradient descent parameters optimization,but its slow convergence speed and the most superior local problems.In order to improve the prediction accuracy of urban road short-term traffic flow,this paper presents a prediction model of WNN IPSO optimization,then the model was applied to the empirical study of short-term traffic flow.The results show that the prediction model of WNN and PSO-WNN with traditional compared,IPSO-WNN model prediction error is smaller,and has faster convergence speed and good ability to nonlinear fitting.3.In view of the traditional financial time series forecasting method is difficult to describe the future trends and changes the space problem,we adopt the IFIG and the SVM prediction model and genetic algorithm(GA)to optimize the SVM,modeling of graining data.The results show that compared with the traditional finncial time series forecasting method can better describe nonlinear dependency relationship between the input and output,so as to more accurately predict the Shanghai index.