Most public buses in our country adopt fixed-point dispatching methods and no combined dispatching mode.They cannot flexibly organize dynamic operations according to the real-time changes of passenger flow,and it is difficult to rationally use traffic resources,save operating costs,and meet the travel needs of passengers.This paper analyzes the related content of bus operation dispatching,and puts forward the research idea of dynamic collaborative bus dispatch based on short-term passenger flow forecasting with fusion of multiple characteristics.The specific research methods are as follows:First,study the collection and preprocessing of bus passenger flow data,analyze the time and space characteristics of bus passenger flow,and determine the prediction method: integrate the three factors of time change law,spatial distribution characteristics and external input characteristics to form a fusion of time and space characteristics and external characteristics of passenger flow.Feature input matrix;then a combined short-term passenger flow prediction model based on CNN and GRU is proposed,which can capture temporal and spatial features separately.Then,it analyzes the dispatching principles,research focus,and application status of full trains,shuttle trains and major station express trains,summarizes their applicable environment and advantages and disadvantages,explores the necessity of combined dispatch,and proposes a joint planning layer and dispatching layer.The multi-mode combined dispatching method forms an over-stop train mode that combines the nature of section operation and cross-station stop restriction,and integrates the whole train in the planning layer.This combination mode is more robust to the uncertainty of bus arrival time and the dynamic characteristics of passenger flow demand,and can enhance the mobility and service efficiency of the bus line network.Next,using the forecast data as input,a multi-mode combination optimization scheme method for public transportation under the environment of the public transportation network is proposed to cover the two-level planning model at the meso and micro levels,while achieving the optimization of scheduling at the time and space levels.Considering the problem of vehicle capacity and passenger detention,making the scheduling model more robust to interference with the actual capacity update of vehicles and changes in passenger flow;starting from the heterogeneity of passenger waiting arrival rates,according to the change between the main and branch lines in the bus network Take the time window to carry out coordinated scheduling between different line departure plans to realize the coordinated scheduling optimization of multiple bus lines,and finally,output a complete network operation plan that includes the departure timetable and the plan of overpassing the station.Finally,through the case analysis of the short-term passenger flow prediction model and the multi-mode dynamic coordinated dispatch strategy of public transportation,the feasibility and practicability of the short-term passenger flow prediction based on the integration of time and space characteristics for the dynamic coordinated dispatch of the line network are verified. |