Data-driven Bus Scheduling Analysis And Optimization

Bus system is in charge of major urban transportation task,and plays an important role in alleviating urban road congestion and abating environmental pollution.However,under the impact of the increasing popularity of private cars and rails,the attractiveness of buses gradually diminishes.As a result,analyzing and optimizing public bus systems has become an imperative issue.The analysis and optimization based on traditional survey data has the disadvantages of long period and weak timeliness.While with the popularization of sensor equipment and positioning technology,a large scale of traffic data is easily accessible.It is of great significance to study how to utilize big traffic data to analyze and resolve the current problem of public bus system in order to improving the attractiveness of public bus system.In response to the problem that massive traffic data in existing researches has been inadequately utilized,this paper combines with taxi track data,bus track data and IC card data to analyze bus scheduling and bus route layout from a novel perspective.In addition,this paper aims at the insufficiency of existing bus vehicle scheduling optimization methods and constructs an optimization model for bus scheduling that considers more constraints.The main research contents and results are as follows:(1)Data preprocessing and calculation.Designing data preprocessing strategies that based on the characteristic of different types of data.Making use of the bus IC card data and bus GPS track data to derive the OD(Origin-Destination)of bus passenger.Extracting taxi passenger OD points from the taxi track data.Determining the bus schedule from the GPS data and calculating the full load rate of each bus line section.(2)Rationality Analysis of Bus Dispatching and Network Layout.Analyzing the problems of public bus system in the study area.This paper proposes the directional unbalanced coefficient of passenger flow and directional unbalanced coefficient of capacity,constructs bus dispatching evaluation model,which considered the coordination between passenger flow and transport capacity,and designs a method based on bus passenger flow data to explore bus routes with irrational scheduling.This paper also discusses the relationship between taxi travel demands and public bus service level.Moreover,the paper proposes a method to discover service blind spots of bus stations based on bus and taxi flow data,and then analyzes the rationality of bus network layout.(3)Bus schedul:ing optimization.Analyzing the deficiency of the existing scheduling optimization model.Building bidirectional scheduling optimization model that considered the constraints of directional unbalanced coefficient of capacity,vehicle resources,and departure time interval and load ratios is built.Using historical bus passenger data and road condition data,combining a genetic algorithm to solve the optimization model.Moreover,verifying the effectiveness of the optimization model through comparing the operational indicators of the public bus system before and after optimization.(4)Visual public bus data analysis system design and implementation.Studying methods and techniques of data visualization analysis,using Openlayers,Echart and D3.js to design the traffic data visualization analysis system.The system implements the following functions:static data display,dynamic data display,bus station service blind spot analysis,optimized schedule analysis and fine-tuning,etc.