Data-driven Customized Bus Route Selection Method

In recent years,customized transit has developed rapidly as a new type of Internet transportation.Compared with conventional buses,customized buses have greatly improved the comfort of public transportation,reduced commuting time of passengers,optimized the allocation of transportation resources,and alleviated the problem of traffic congestion to a certain extent,reducing the amount of pollution emissions from society.However,according to the survey,at this stage,many enterprises plan the route manually,which leads to unreasonable route planning,reduces the quality of customized bus service,and reduces the attractiveness of the product to users.In order to solve this kind of phenomenon,this paper studies the method of customized bus route planning.This article first elaborates the concept,characteristics and operation process of the customized bus system,and analyzes some problems in the current operation of the customized bus.Based on these problems,this paper proposes a data-driven customized bus route planning method.This method uses data mining,operations research and other knowledge to plan stations and routes.It is more scientific,accurate and reasonable than traditional manual route planning.The route planning method in this paper is divided into three steps: site selection,model construction,and model solution.In terms of site selection,first of all,the travel data is collected and preprocessed.Secondly,this paper compares two more commonly used clustering methods,and finally uses the k-means method as a method of site clustering,and analyzes the deficiencies of k-means.At this point,the clustering results are planned through the contour coefficient and the visual evaluation of the stations to plan a reasonable boarding and alighting station.Secondly,the custom bus route planning model is constructed in combination with the vehicle routing problem model.The custom bus route model is designed from the perspectives of society,passengers,and enterprises,and the multi-objective combination optimization function with the greatest social benefits and the highest corporate profits is established,combined with customization Relevant constraints have been added to the actual operation of the bus and passenger ride experience,making it more in line with the current operating status and customized bus characteristics.The genetic algorithm is used to solve the model problem,and the method of finding the shortest path based on optimal segmentation improves the corresponding coding rules.Finally,through the actual travel data of Shenzhen users as a case analysis,several customized bus lines were designed,and 7 lines were selected for analysis.The results show that the model and algorithm can plan a more reasonable customized bus planning plan,which has reference value and research significance for enterprises and operators.