Short-term Traffic Flow Prediction Based On ABOA-LSSVM

With the development and progress of the current society,the problems of the traditional traffic method theory in practical application continue to highlight.At this time,intelligent transportation system,as the integration of advanced science and technology and transportation theory,will play a more and more important role in solving road congestion,ensuring travel safety and reducing environmental pollution.Accurate and efficient traffic flow prediction in intelligent transportation system not only helps to achieve real-time traffic control,but also helps to provide some related applications for active traffic guidance,so as to provide guarantee for later travel.Therefore,the research in this area has been favored by a large number of relevant researchers,which has very important practical significance.This paper summarizes the existing short-term traffic flow prediction methods and models.Combined with the randomness,nonlinearity and spatiotemporal correlation of short-term traffic flow data,in order to meet the prediction requirements,it can be combined with practical application.The least squares support vector machine(LSSVM),which is suitable for solving nonlinear problems,has high speed and good portability,is selected to carry out short-term traffic flow regression prediction.For the difficulty of parameter selection in LSSVM regression prediction model,a simple and efficient whale optimization algorithm(WOA)is used to optimize the parameters.And the WOA is improved,get an optimized algorithm(ABOA).Then the traffic flow prediction model which is constructed by combining the two methods is used to forecast the traffic flow data of the processed expressway,the main contents of this paper are as follows.(1)This paper collects the traffic flow data of K2077 section of Shankun Expressway in Yunnan Province,and the traffic flow of five working days from May 7 to May 11,2018 is selected for the experiment.Firstly,considering that the traffic flow data obtained on the road is easily interfered by noise data and missing data,which affects the prediction accuracy,the data needs to be repaired;Finally,the effective traffic flow data is obtained by wavelet threshold de-noising method,which provides a basis for ensuring the accuracy of short-term traffic flow prediction.(2)The convergence speed of WOA is not fast enough in the early stage,and it is easy to fall into the local optimal solution in the late stage of iteration.An improved whale optimization algorithm-ABOA,is proposed to improve WOA by introducing adaptive weight and beetle antennae search.In the improved algorithm,the search for prey strategy in the original WOA algorithm is replaced by the beetle antennae search strategy,which speed up the early search ability and improves the search accuracy;At the same time,the adaptive weight is introduced into the encircling prey and bubble net attack stage to enhance the local search ability in the later stage of iteration.The effectiveness of the improved strategy is proved by using three kinds of 18 different standard test function cases.(3)This paper combined with the advantages of fast training speed,strong global and high accuracy of LSSVM.It is selected as the basic model of short-term traffic flow prediction.At the same time,in order to further improve the prediction accuracy,the values of penalty parameter and kernel function parameter in the model are selected appropriately.The improved algorithm is used to find the optimal parameters of LSSVM,the short-term traffic flow prediction model based on ABOA-LSSVM is constructed.The optimized traffic flow data is divided into test group and training group for experiment,and the performance of the model is judged by comparing the prediction results with the actual value.Through compares the evaluation indexes of the prediction results of ABOA-LSSVM,WOA-LSSVM,PSO-LSSVM,LSSVM and BPNN(BP neural network).The results show that the prediction accuracy of LSSVM optimized by ABOA is higher,and it is proved that the proposed aboa-lssvm model is effective in short-term traffic flow prediction.