Research On Spatial Distribution Evolution And Rebalancing Strategy Of Dockless Shared Bicycles

With the development of sharing economy,free-floating shared bicycles have gradually emerged and become another green travel mode after the docked public bicycles.At the same time,the “tidal phenomenon” of the free-floating bike-sharing system has also caused travelers to face the problem of no bikes to borrow.Therefore,rebalancing bikes scientifically and reasonably has become a significant issue.In this paper,the actual cycling data of Qingju bicycles in Nanjing are studied.It is found that the spatial evolution of shared bicycles is relatively stable on the monthly time scale and the movement of bicycles in space is approximately linearly accumulated over time.We found that the collective motion of shared bikes has significant anisotropy and discussed the mechanism behind the anisotropic motion.The supplydemand distribution of shared bicycles presents the spatial characteristics of “large imbalance concentrated in a few areas in the center while small imbalance dispersed in the periphery”,which provides us with a new idea of designing a more efficient rebalancing strategy.In addition,the results of data mining show that rebalancing behavior of operators tends to restore the initial spatial state of shared bikes.Based on the evolution law of the spatial distribution of shared bikes,this paper designs a combined rebalancing strategy in which operators and users participate together.Operators’ truck rebalancing can solve the problem of bike imbalance in urban central areas.User rebalancing is that operator uses a negative price strategy to encourage cyclists to ride “red packet bikes” from areas with an oversupply of bicycles to areas with an undersupply to rebalance bicycles in marginal areas.With the goal of minimizing the total repositioning cost,we use mixed integer programming to formulate the combined rebalancing strategy,and different strategies are compared in a small-scale example.The results show that the cost of combined rebalancing is lower than that of single repositioning.In addition,the effects of negative price coefficient and unit penalty coefficient in the model on the strategy is further discussed.Finally,this paper applies the combined bike-sharing rebalancing optimization model to the real-world scenario of Nanjing Xianlin University Town for analysis.Taking the operator rebalancing volume obtained in the previous article as the expected demand volume of the site,a genetic algorithm with local search operation is designed to solve the repositioning strategy and route of the combined bike-sharing rebalancing optimization model in the real-world scenario.The results show that the combined rebalancing strategy can be well applied to the actual scenario of shared bike scheduling.Besides,it can provide theoretical guidance on the actual rebalancing for operators.