Multi-Source Data-Driven Optimization Of Customized Bus Routes In New Urban Areas

The rail transportation in the new urban areas is still in its initial stage,and the rail transportation network has not been perfected into a network.Regular buses are generally the main force of public transportation in the new urban areas,taking up a large proportion of the public transportation service.Most of the new urban areas are in the rapid development stage,and there are a large number of commuting groups in the morning and evening peaks on weekdays,with low commuting efficiency,long commuting time,and the need for multiple transfers.Based on this background,customized public transportation was born.Customized bus routes can provide commuting groups with faster,more comfortable and personalized travel services,which can optimize the allocation of transportation resources and improve the attractiveness of public transportation;at the same time,reduce urban road traffic congestion and air pollution problems.By mining multi-source data of public transportation,analyzing the spatial and temporal characteristics of public transportation trips and commuting characteristics,and studying custom bus stop location and route optimization based on commuter flow data and workplace distribution,we can provide custom bus services that can better meet the travel needs of commuting groups.Based on this,the main research of this paper is as follows:Firstly,based on the bus multi-source data,the identification method of boarding and alighting stations,transfer trips and commuter trips based on travel chain is proposed to get the bus travel data and bus travel characteristics.Based on the temporal and spatial dimensions,we analyze the characteristics of bus trips.From the perspective of time slicing,we analyze the time-varying characteristics of passenger flow on weekdays and rest days,and at different times of the day;from the perspective of spatial distribution,we analyze the hotspots of bus trip activities on weekdays and rest days,and analyze the relationship between bus trip hotspots and urban space;from the perspective of commuting pattern,we analyze the spatial distribution of commuting trip hotspots and commuting distance characteristics,and identify the morning and evening peak commuting From the perspective of commuting pattern,we analyze the spatial distribution of commuting hotspots and commuting distance characteristics,identify the hotspots of occupational and residential distribution of commuting in the morning and evening peaks,and dig out the commuting pattern of public transportation to provide data support for site selection and route optimization.Second,on the basis of acquiring the travel hotspot areas and commuter travel demand OD data,an improved passenger flow weighted K-means clustering algorithm is proposed to consider the influence of commuter traffic carried by bus stops on the total distance traveled between stops,while the contour coefficient method is used to optimize the K-means clustering algorithm to study the siting of carpool stops.Comparing with the traditional K-means clustering results,it is found that the improved passenger flow weighted K-means clustering center is closer to the commuter flow gathering area,and the effectiveness of the improved Kmeans clustering is further verified by taking different values such as 4,5,6 and 7,which provides a more scientific and reasonable method for custom bus stop location selection.Finally,based on the bus passenger flow commuting characteristics and the results of carpool site selection,a customized bus route optimization model with the optimization objectives of travel cost of travelers and minimum bus operation cost was constructed considering the realistic scenario of multiple bus stations in multiple areas in Xi’an New Area,and the model was solved using genetic algorithm and verified using Xi’an New Area as a real case.The model and algorithm were found to meet the requirements of custom bus planning and can be applied to the actual scenario.Compared with the conventional bus,the custom bus can save 37.89% of the average commuting time,which can provide fast and comfortable bus travel for the commuters in the new urban area and fill the gap of the current urban rail transportation.The model and algorithm are verified to have good practicality and effectiveness,providing theoretical support and practical application value for the study of customized bus route optimization.