Research On Prediction And Evaluation Method Of Urban Road Traffic Congestion Status

With the rapid growth of the social economy,urban traffic congestion has become more and more serious.Traffic congestion has affected the work and study of the majority of residents,accompanied by severe environmental problems.Therefore,it is especially important to alleviate the traffic congestion problem.If residents can make timely and accurate predictions and evaluations of traffic congestion that has occurred or is about to occur in the future,it can effectively avoid the impact of traffic congestion.Aiming at the insufficiency of the existing traffic congestion prediction methods,such as low precision,weak stability,long-term prediction time,and the different impacts of traffic time on traffic congestion,and the single factor can not accurately represent the traffic congestion state,this paper proposes a method for predicting and evaluating traffic congestion status levels using multi-index fuzzy comprehensive evaluation method.In the aspect of prediction,a short-term traffic flow prediction model based on random inertial weight particle swarm optimization combined with support vector regression machine(MPSO-SVR)was proposed,which is used to predict the average speed and traffic flow of traffic flow parameters that characterize traffic congestion.In this model,the uniform distributed random inertial weight are substituted for the invariant inertial weight of the standard particle swarm algorithm,which accelerates the convergence speed of the algorithm and improves the performance of the algorithm.In the aspect of traffic congestion state evaluation,the predicted values of the three factors: the average speed,the traffic flow density,and the road saturation are obtained from the predicted values of traffic flow and average speed,and input the three factor indexes to Multi-index fuzzy comprehensive evaluation model that establish the set of factors and evaluation(level)for traffic congestion.It determined the weight coefficients of three factors under the morning peak and the evening peak and other periods by the entropy method,then determined the degree of membership of each index in each period by the trapezoidal membership function.Finally,it divided the traffic congestion state into six levels.Finally,through the traffic congestion prediction evaluation experiment on the trafficdata of the I405 highway in the US PeMS database,the predicted average speed and traffic flow are calculated to obtain the traffic congestion index system,and used in the multi-index fuzzy comprehensive evaluation method to predict the traffic congestion level,and compared with the traffic congestion level evaluation results predicted by the standard PSO-SVR method,it can be seen that the traffic congestion state predicted by the optimization method proposed in this paper is in good agreement with the actual state,and the accuracy rate can reach 94.79%.