Optimization Of Shared Bicycle Dispatching Path Based On Inventory Rebalance

Recent years,the rapid development of the shared bicycle pattern becomes a solution to the last mile problems.Due to its pile-free design,and the use of the Internet and smart devices,this pattern solves the inconvenience of the government-led public bicycles,and truly embodied the convenience of the bicycle.However,shared bicycles are a resource with mobility.Residents lead to the relocation of bicycles,and at the same time causes the flow of resources in space,resulting in imbalance of limited resources.In other word,the movement of the shared bicycle will lead to the spatial-distribution imbalance of bicycles,and the residents may encounter the situation that there is no bike to ride.Therefore,shared bicycle operators need to ensure that resource is balanced to increase customer satisfaction and minimize operating costs.The demand forecasting of bicycles and the subsequent research on the distribution plan of bicycle dispatching respectively correspond to the problem of inventory balance and operation cost,which has strong practical significance.In this paper,a demand forecasting method based on improved random forest is proposed.By pre-screening the training data,similar data with higher quality that is beneficial to prediction is selected as the training data set.In addition,based on the availability of data,this paper proposes a series of factor system that may affect the demand of shared bicycles.Through the machine learning mechanism of random forests,the inherent hidden pattern between the factors is mined to ensure high prediction accuracy(MAPE 8.2%).Sensitivity analysis method is also applied to test the influence of parameters on the prediction results.Strong prediction of tree depth can improve prediction.The size of forest can be adjusted according to the scale of the original available data.Compared with the traditional random forest prediction model,the proposed forecasting model based on improved random forest performs better on the variance of the prediction results.On the issue of rebalancing of shared bicycle resources,this paper quantitatively describes the imbalance between supply and demand of bicycle resources,considering the demand,supply and the number of damaged bicycles in the region,and proposes a quantitative description of the spatial distribution of bicycle resources,which is the bicycle resource.Rebalancing scheduling provides early theoretical and data support.The equilibrium description proposed in this paper is based on the idea of relative balance.The starting point is to measure the relative fairness of resources in spatial distribution,rather than a simple balance of supply and demand.In order to simplify the distribution description of resources and the location distribution of bicycle placement points,this paper adopts the common gridding method to mesh the research area,which not only conforms to the actual operation mode of the enterprise,but also simplifies the difficulty of model solving and facilitates the analysis of results..In the problem of vehicle distribution path caused by shared bicycle rebalancing,this paper identifies the actual type of the problem,that is,there are problems of simultaneous collection and departure of bicycle placement points.Therefore,this paper models the problem of simultaneous transmission and reception of vehicle paths,and designs a corresponding heuristic algorithm to solve it conveniently.The partiality of Beijing is taken as the research object to verify the effectiveness and stability of the algorithm.Based on the case analysis,the problems arising from the operation of the shared bicycle business are analyzed and explained,involving bicycle delivery and daily dispatch.