Research On Optimal Scheduling And Deployment Of Unmanned Taxi Based On Trip Demand

With the significant improvement in the living standards of our residents,people are increasingly demanding the comfort and convenience of trip,the number of private cars has risen sharply,and traffic congestion has become increasingly serious.There is a contradiction between residents' growing personalized trip demand and limited taxi resources.In addition,the game between drivers,passengers and operating platforms has intensified the contradiction between supply and demand.In this context,in order to meet the residents' demand for taxis,simply increasing the number of taxis not only fails to completely solve the problem of taxis,but also wastes limited road resources and aggravates urban traffic congestion.Only through the unscheduled taxi resource optimization scheduling and deployment,effectively improve the taxi utilization rate and reduce the no-load rate,can alleviate the urban traffic congestion problem while meeting the trip demand.In this paper,under the premise of meeting all trip needs,the minimum number of taxis required and the reduction of the no-load rate as the main optimization goal,around the trip demand matching and trip demand forecast,considering the characteristics of unmanned taxis and reasonable Iterative connection time,guide the unmanned taxi resource allocation,vehicle deployment and dispatch through match demand matching,in order to alleviate the contradiction between supply and demand of taxis,guide passengers to choose public transportation,ease urban traffic congestion,and self-driving taxi in the future.The promotion and application provide a theoretical basis.The details are as follows:1.City taxis get on and off the hotspot area identification.Through the Mean-shift clustering algorithm,the hotspot area of the city taxi is identified.Analyze the characteristics of unmanned taxis,propose the operation mode of fixed and drop-off points for taxis,and determine the cluster center of hotspots to adjust the fixed passengers to the appropriate location.2.Trip demand matching and vehicle scheduling research.Through the taxi trajectory data between the hotspots in Suzhou,use the bipartite graph to establish a matching model for the inter-regional demand,establish a demand matching model,and use the least vehicle to complete all trip demand.The problem is expressed and converted to a minimum with a clear formula.Path coverage problem,using Kuhn-munkras algorithm to obtain the required number of taxis,reduce the no-load rate of taxis and load balance the operating vehicles.3.Short-term trip demand forecasting and vehicle deployment studies.Demand matching relies on the trip demand submitted by passengers to dispatch vehicles.It is difficult to respond quickly to temporary trip demand,short-term trip demand forecasting,to deploy vehicles ahead of time,shorten the lead time for passengers to submit demand,and reduce passenger waiting time.Firstly,the characteristics of trip demand data and the characteristics of time and space change are analyzed.LSTM is used to predict the trip demand of different regions in different time periods.Based on the trip demand matching model,a dynamic demand matching model is established to provide the deployment and dispatch of unmanned taxis.Theoretical guidance.The final result proves that the taxi resource scheduling and deployment plan obtained through the two-tier demand matching model can reduce the operating vehicle by 35% and reduce the no-load rate to 31.4% based on the completion of all trip demand.The population density is large,the trip demand density is large,and the urban areas with more uniform distribution in residential areas,industrial areas and commercial areas are better.