Research On Rebalancing Problem Of Free-floating Bicycle Sharing

As an innovation of private bicycles,station-based bike sharing(SBBS)and traditional bicycle rental models,Free-floating bike sharing(FFBS)are more convenient and flexible with their own public and “no stations” characteristics.Once it came out,it was sought after by users and capital.However,with its in-depth popularity,its imbalance between supply and demand has become more and more prominent,which has seriously affected the user's personal experience,the interests of enterprises and the order of society.Therefore,how to formulate a more effective rebalancing plan and truly open up people's last mile travels have become the focus of attention from society and are also the main research content of this paper.In this paper,FFBS rebalancing is divided into static rebalancing which occurs with little user interference at night,and dynamic rebalancing which occurs with much user interference during the day.For static rebalancing,this paper considers the characteristics of static complete rebalancing in the actual operation process,and establishes a split vehicle routing problem model with simultaneous delivery and pick-up for the goal of minimum total rebalancing cost and minimum rebalancing completion time.Then this paper proposes a hybrid algorithm based on network decomposition,greedy algorithm and variable neighborhood search to solve the problem.In order to make the example adapt to the model and algorithm,the K-means algorithm is used to process and analyze the data in the example.Finally,the improved algorithm can obtain better solution than traditional variable neighborhood search.The results show that the proposed model and algorithm are effective,and can provide theoretical basis and practical guidance for static complete rebalancing problem of large-scale and multiple distribution vehicles.For dynamic rebalancing,this paper considers its characteristics in the actual operation process,then uses the rolling horizons to transform dynamic rebalancing into a series of static rebalancing(ie,static rebalancing in the initial stage and the dynamic stage)to respond to use interference.Contrapose their relationship and difference,this paper establishes a split vehicle routing problem model with simultaneous delivery and pick-up and soft time window for the goal of minimum total rebalancing cost and maximum users' satisfaction,which also considers the influence of the impact of congestion on some sections of the commuter peak period.Then design a hybrid algorithm with adaptive substructure based on the third chapter of this paper to make it adaptable to dynamic rebalancing.Then the long-short term memory and linear neural network are used to model the case data to obtain the gap between user interference and forecast.Finally,the effectiveness and advancement of the model and algorithm is illustrated by the example and the impact of congestion on the rebalancing scheme is analyzed.