Research On Data Driven Facility Location And Repositioning Optimization Of Free-floating Bike-sharing System

As a low-carbon,low-cost,convenient,and sustainable mode of urban public transportation,free-floating bike-sharing system has been favored by governments and people around the world for its unique product advantages.However,due to the uncertainty,tidal nature,and improper operation of user travel demand,the problems of supply and demand imbalance,disorderly parking,and bike damage are very prominent in the free-floating bike-sharing system,seriously affecting the operation order of urban public transportation and people’s daily travel experience.Therefore,scientifically planned free-floating bike-sharing infrastructure is a key link in improving the efficiency of urban public transportation services,and an efficiently operated free-floating bike-sharing system is a basic guarantee for enhancing people’s daily well-being,with significant importance.This study focuses on the optimization of facility location and repositioning optimization for free-floating bike-sharing systems from two aspects: strategic planning,and operations management.At the strategic planning level,this study investigates the location optimization problem of electronic fences and repair stations for free-floating bike-sharing system.At the operations management level,this study investigates the optimization of repositioning between areas and within areas for free-floating bike-sharing system.To address these issues,different mathematical optimization models were constructed from different perspectives.Corresponding solution methods were designed and experiments were conducted based on actual operating data.This study mainly focuses on the following four aspects:(1)This study investigated the location optimization problem of electronic fences for free-floating bike-sharing systems considering the coordination with roadside parking spaces.To address the uncertainty and periodicity of the parking demands in free-floating bike sharing system,a two-stage stochastic programming model based on user parking behavior was proposed,and a scenario-based stochastic programming model as well.In order to solve the proposed model,a genetic algorithm based on sampling approximation was designed.The results show that the genetic algorithm based on sampling approximation can quickly converge when solving this problem,and its solving ability is superior to traditional genetic algorithms.(2)This study investigated the location optimization problem of repair stations for free-floating bike-sharing systems considering the waiting time in the queue.To address the issue of stacking of broken bikes in repair stations,a repair station coverage location model based on queuing theory was constructed.A quadratic programming method was proposed to transform the high-order constraints in the repair station coverage location model into a form that can be accurately solved.The advantages and solution scale of the model were verified,and the role of important parameters in the model was analyzed.The experimental results show that the coverage site selection model based on queuing theory can shorten the average transportation distance of damaged bicycles by 23.37%,and reduce the service queuing time in damaged bicycle stations by 157%.(3)This study investigated the inter-area repositioning optimization problem of free-floating bike-sharing system considering multiple types of transportation vehicles.To address the issue of supply and demand imbalance between different areas of free-floating bike-sharing system,a basic optimization model for inter-area repositioning of free-floating bike-sharing system considering multiple types of transportation vehicles was constructed,the properties of which were proven.A chance-constrained model based on the basic model was proposed,and an improved genetic algorithm considering multiple evolution mechanisms was designed.The effectiveness of the model and algorithm was verified.The experimental results show that the block crossover strategy can speed up the convergence speed of the algorithm,and the user satisfaction level of the chance constrained programming is 13.4%higher than that of the mean model.(4)This study investigated the optimization of intra-area repositioning problem of free-floating bike-sharing system considering multiple repositioning strategies.To address the diversity of repositioning scenarios in free-floating bike-sharing system,three repositioning strategies as well as corresponding mathematical models were proposed.A neighborhood search algorithm based on adaptive large neighborhood and variable neighborhood nested was proposed,and the effectiveness of the model and algorithm was verified.An optimal strategy selection method for different repositioning scenario was provided.The experimental results show that when the dispersion of handling scenes in the area is less than 0.5,the mixed handling strategy is better,and when the dispersion of handling scenes in the area is greater than 0.5,the action-separated handling strategy is better.The research findings of this paper can provide theoretical basis for the government in the strategic planning of urban public transportation facilities,and have a guiding role in improving the operations management and service efficiency of free-floating bike-sharing companies.