Analysis Of Bus Operation Characteristics Based On GA-ARIMA Model

With the rapid development of cities,the number of cars is increasing,and the problem of road congestion has become a hot issue currently facing.As an important part of the road traffic system,the public transit system has become an effective force for alleviating traffic jams.At present,buses in many cities have been equipped with on-board GPS positioning equipment,which is conducive to collecting operational data of buses and can monitor the real-time status of buses in real time.In addition,the collected driving data of the bus can effectively analyze and predict the running characteristics of the bus.Not only can the forecast information of the bus be released to improve the passenger’s riding satisfaction,but also can be carried out better.Issues such as scheduling and emergency response of bus vehicles can effectively improve the utilization rate and passenger capacity of buses.The main research contents of this article are as follows:Firstly,the paper preprocesses the GPS data collected by the bus positioning equipment through the data processing methods such as abnormal data processing,data compensation,map matching,geometric location distance calculation,etc.;using the processed effective data to analyze the bus operation characteristics such as the travel time of the bus,the station in and out time,and then the genetic algorithm and time sequence The ARIMA optimization model based on genetic algorithm is established.Through the model optimization,the prediction method of bus travel time and other operation characteristics is proposed,and the prediction results are evaluated and analyzed through the robustness and other error evaluation indexes.Through the actual case analysis,the error evaluation indexes of the prediction results obtained by using ARIMA model are:robustness=19.71%;MAPE=14.37%;RMSE=12.93%,the error evaluation indexes of ga-arima model are:robustness=17.44%;MAPE=12.33%;RMSE=10.67%.The evaluation results of the two models are compared and analyzed.The results show that the data results and prediction accuracy are improved to a certain extent after optimizing the parameters of time series model by genetic algorithm.The feasibility of the model in predicting bus travel time and other operating characteristics can be applied in practice.