| With the development of diversified public transport,customized bus has been active in the public eyes and gradually become a major research hotspot.However,at this stage,most of the customized bus operation lines are mostly morning and evening peak commuter lines,with single operation mode and limited service scope.The collected passenger travel demand is only the commuter demand from residence to work place,ignoring the difference of travel demand.In addition,the network planning of customized bus now is more based on the reference of conventional public transport,the planning method does not match its own characteristics.Most of the operating enterprises are facing losses,and the popularity of customized public transport is not high,so it is difficult to get better development.In order to further expand the influence of customized bus and promote the transfer of private transport to public transport,it is necessary to conduct in-depth research on the characteristics of passenger travel demand and network planning of customized public transport.Based on the research status at home and abroad and the actual operation status of customized public transport,this paper first analyzes the connotation and characteristics of customized public transport,and defines the function orientation and key service elements of it.Then,starting from the characteristics of passenger travel demand,the paper proposes that there are differences in passenger travel demand,and further divides passenger travel demand into general commuter travel demand and diversified travel demand according to passenger types.According to the different characteristics of passenger travel demand,such as different travel time,scattered starting and ending points,and diversified types of passengers,this paper proposes a method of space-time two-dimensional clustering to deal with the spatial and temporal attributes of combined travel demand.The improved k-means algorithm is used to cluster the spatial location of combined travel stations,and the hierarchical clustering method is used to cluster the travel time.Combined with the characteristics of customized public transport and travel demand,this paper puts forward the corresponding principles,processes,basic contents and methods of network planning.For the relatively constant commute travel needs,this paper selects the typical mode of multi-departure to single-destination for research,and establishes a network planning model with the minimum total cost and the maximum network coverage as the optimization objectives,and the above occupancy rate,number of stations,vehicle capacity limit and the latest arrival time as constraints.For the diversified travel demand,in addition to considering the maximum profit of the operating enterprise and the maximum coverage of the network,this paper also takes the minimum total detour distance of the network into the objective function,considers the travel time window constraint in the constraint conditions,and considers the multiple models in the model selection.In the design of ant colony algorithm,the difference of travel demand is considered,and the customized bus network of various modes can be obtained according to the actual situation of passengers’ travel demand.Finally,for the fixed planning model of commuter demand line network,an example is designed to take Dongfanghong square of Lanzhou City as the destination.For the planning model of diversified travel demand line network,take the travel demand of Anning District,Qilihe district and Chengguan District of Lanzhou City as examples.The validity of the model and algorithm is verified by comparing and analyzing the results of route planning.The results show that the customized bus network planning method proposed in this paper can effectively meet the different travel demands of passengers,bring benefits to the operation enterprises,expand the influence of customized public transportation,and provide certain theoretical basis and support for the actual operation of customized public transportation. |