Research On Regional Bus Scheduling Based On Passenger Flow Forecast

As the contradiction between supply and demand of traffic becomes increasingly prominent,urban traffic congestion and other problems need to be solved urgently.Giving priority to the development of public transportation is a good way to solve the problem of traffic congestion.As one of the most commonly used modes of public transportation,bus has been widely used in cities across the country.It has the characteristics of low investment cost and high accessibility,and there is no need to build special lines.In the bus operation planning,the formulation of schedule and driving plan is a very important part,reasonable schedule and driving plan can improve the bus service level and reduce the operating cost.However,at present,there are still some cities in the timetable and the formulation of traffic plans in unreasonable circumstances,contrary to the concept of giving priority to the development of public transport.For this,this article is based on historical data to predict the future traffic demand,passenger flow according to the traffic demand changes to develop dynamic schedule meet the needs of passengers,and according to the schedule to build fuel bus and pure electric bus regional traffic plan,the city’s public transport vehicle scheduling problem is studied,the main work content is as follows:(1)In order to facilitate the prediction of bus passenger flow data in the later stage,it is necessary to conduct correlation analysis on multi-source bus data.In this paper,Pearson coefficient is used to analyze the correlation between the data.The correlation between weather attributes and passenger flow is analyzed,and the weather attributes that affect passenger flow are extracted.The correlation between the historical passenger flow and the current passenger flow under the time dimensions of hour,day and week was analyzed,and the historical passenger flow was taken as the model input through comparative tests,which effectively improved the accuracy of the model prediction.(2)On decision tree,LSTM and GRU helped model of public transport passenger flow data to predict the effect not beautiful problem,this paper designed a model based on the LR-GRU helped GRU helped,the model USES LR module training the external factors influencing the passenger flow,using GRU helped module training history traffic sequence,effective use of the factors influencing the passenger flow external and internal passenger flow time series characteristics,Compared with decision tree,LSTM and GRU model,LR-GRU model improves the accuracy of bus passenger flow data prediction.(3)In this paper,a regional bus scheduling optimization model is designed,which is composed of the upper-level planning model of scheduling schedule and the lower-level planning model of mixed regional scheduling.The upper model aims at minimizing the waiting cost of passengers and the total mileage cost of bus line operation,and is constrained by the departure frequency.The lower model aims to minimize the operating cost,and is constrained by the maximum vehicle mileage,task time,station capacity and pollution gas emission.Compared with the static line schedule and driving plan of the original bus operation plan,this paper designs the optimization model of regional bus vehicle scheduling to work out the line schedule and regional driving plan that changes dynamically with the passenger flow,which not only improves the level of bus service but also reduces the operating cost of bus enterprises.(4)Aiming at the strong constraint problem of the lower level programming model,this paper designs the coding scheme,the initial population generation algorithm,the crossover operator and the mutation operator to improve the standard genetic algorithm,and solves the problem that the standard genetic algorithm can easily produce infeasible individuals when solving the strong constraint problem.In order to verify the effectiveness and practicability of the above method,this paper gives the specific experimental design process,and verifies the effect of the method proposed in this paper through experiments,gives the test results,and analyzes the test.The experimental results show that the method designed in this paper can solve the problems of unbalanced supply and demand between bus enterprises and passengers and unreasonable planning to a certain extent.