Study On Static Demand-Responsive Public Transportation Route Planning Based On Potential Passenger Flow

With the rapid development of economy,urban traffic congestion has been a key issue restricting the further development of cities.Promoting travel by public transport is one of the measures to alleviate urban traffic congestion.However,at present,there are insufficient investments in ground public transport systems in many cities,and the development of ground public transport systems is not able to keep pace with the growing demand of urban residents for diverse and personalized travel.As a new auxiliary public transportation mode,demandresponsive public transportation system is able to enrich the service mode of ground public transport systems and meets travel demand of urban residents for diversification and personalization.Therefore,further research on route planning methods for demand-responsive public transportation system is practical significant to develop traditional ground public transport systems.Firstly,relevant concepts of demand-responsive public transportation are expounded,and characteristics and classification of demand-responsive public transportation are introduced.Static demand-responsive public transportation is determined as the research object.By analyzing relationship between demand-responsive public transportation and other transportation modes and comparing their characteristics,the advantages of demand-responsive public transportation are obtained.It is possible for passengers to shift from other modes of transportation to demand-responsive public transportation.Secondly,based on the passenger characteristics of static demand-responsive public transportation,identification criteria for potential passenger flow are established.The criteria are used to identify and mine potential passenger flow of static demand-responsive public transportation from taxi GPS data.Passenger OD flow is defined from the perspective of spatial geography and geometric characteristics.By analyzing the similarity between OD flow from three aspects(spatial geography,travel distance,and travel angle),a similarity function is developed.OD flow clusters are identified and mined from taxi GPS data by using the KNNaggregation hierarchical clustering framework.Potential passenger flow is identified by the criteria.Thirdly,a static demand-responsive public transportation stops location model is constructed.Considering the service range and service coverage of stops,the objective function is to minimize walking distance between passengers and the stops and to minimize the number of stops.An adaptive k-medoids algorithm is used to solve the model.Fourthly,considering passenger time window,a static demand-responsive public transportation route planning model with multi-objective is constructed,which takes the benefit of both passengers and operators into consideration at the same time.The objective function is to minimize the cost of passenger travel time and to minimize the total cost of service.To optimize this system,a non-dominant sorting genetic algorithm(NSGA-2)with forward insertion algorithm is developed.Virtual stops are created to solve the constraint that traditional Vehicle Routing Problem(VRP)is not bale to serve the same stop,which makes the model more real and reliable.Finally,to apply the method of potential passenger identification,stop location and route planning for static demand-responsive public transportation proposed in this paper,Xi ’an is selected as a case study,The result shows that the method is able to provide several planning schemes whose both the cost of passenger travel time and the total cost of service are small to decision-makers.All schemes can effectively serve potential passengers,and reduce the travel cost of passengers and the total cost of operators.Reducing the occupancy of urban road resources,urban traffic congestion alleviate.