| With the development of economy and the acceleration of urbanization,the urban population continues to increase,and the scope and magnitude of the urban population mobility are also increasing,thus travel issues have become a hot spot in the current social research.As a mainstream mode of travel,public transportation plays an important role in alleviating traffic pressure and improving residents’ travel experience.However,due to the shortcomings such as transfers,schedules,and fixed routes,buses are becoming less attractive to the residents to travel.Meanwhile,the ratio of taxi and private car travel has been increasing in recent years,which brings higher pressure on traffic management.Hence,optimizing bus routes and increasing the attractiveness of bus travel are effective means to relieve the traffic pressure and enhance residents’ travel experience.However,traditional bus route design schemes often only pay attention to the distribution of passengers at stations along the route,while ignoring the OD(Origin and Destination)of passenger travel and the bus carrying capacity.This paper has proposed a bus route optimization scheme based on the taxicabs trajectory data and GCN(Graph Convolutional Network)technology.The scheme has specifically analyzed the problems of current bus routes and bus stations in update lagging,crowded,unreasonable route,etc.,in combination with the methods of station update analysis,travel flow analysis,GCN,etc.,it has optimized the bus route to make it better adapts to the travel demands of the current resident.The main content and results are:1.Analysis of Station Update demand:This paper has analyzed the current problems of unreasonable bus stop location,update lagging and so on,it carried out the DBSCAN(Density-Based Spatial Clustering of Applications with Noise)analysis based on the actual travel distance on the the location of the surrounding taxicab boardings and alightings,POIs location,etc.,compared and analyzed the cluster centriods which were projectd onto the road network with the real stations,thus to select the most suitable location for the station.Finally,three bus stops with too small spacing in the 203 bus line were deleted,and the locations of five bus stops,such as Diwang Building and Cuizhu Building,were optimized.2.Analysis of Travel Flow:In combination with the taxicab trajectory data,subway card tapping data,etc.,it has analyzed the OD of residents’ travel,which truly reflected the residents’ travel demand.When optimizing the bus routes,it has taken the changes in travel hotspots caused by station selection into consideration,so as to avoid the over or too little capacity of the bus route.After analysis,it is found that the main hotness of 203 bus line is concentrated between Dongmen Station and Shanghai Hotel East Station.In addition,the area around the Huanggang Port area and the railway station are closely connected with the 203 bus line.3.Route Optimization Scheme Based on GCN:This scheme has designed a two-layer graph network,performed the convolution operations respectively in terms of the travel flow heat and the actual road network,and obtained the weight of the stations as the basis for the optimization of the bus route.Meanwhile,it has also adopted the elliptical pruning algorithm to reduce the search range of nodes and improve the computing efficiency.Through comparison,it can be found that the optimized bus route is better than the original bus route in non-linear coefficient,coverage degree,total hotness and other indicators,which proves the effectiveness of GCN method in bus route optimization.In the end,this paper has combined the taxicab data and road network data of Futian and Luohu District in Shenzhen to optimize a local bus route,and compared and analyzed the actual bus route and the results of other bus route optimization schemes.The analysis result has shown that the scheme proposed in this paper has achieved the purpose of adapting to residents’ travel demands and improving residents’ travel experience. |
PDF Full Text Request