Study On Scheduling Optimization Of Urban Public Bicycle System

Sustainable development is the premise of economic construction,and the development of urban transportation is facing the challenges of resource consumption and environmental pollution.Green transportation aiming at low energy consumption and pollution is rising,and low-carbon travel has become a global consensus.As a short-distance transport,public bicycle plays an important role in energy saving and emission reduction while providing convenience for users.The emergence of dockless public bicycle system has a great impact on short-distance travel in the city.Whether the traditional docking system or the new type dockless system,the difficulty of borrowing or returning bicycles has always been a key factor restricting the development of the system Optimizing the scheduling service of the system is one solution to the problem.Based on the analysis of the operation characteristics of the system,this paper studies the scheduling optimization of two types of systems with the objective of reducing the operation cost of these enterprises.Below is the list of main works and relevant conclusions:Understanding operation patterns of public bicycle system.According to the operation data of the system,understanding the operation patterns of the system is the basis of Data-Driven Decision Support(DDDS).Taking Boston public bicycle system as an example,this paper tries to understand the operation characteristics of the system through qualitative and quantitative analysis from six aspects:the basic spatial-temporal distribution of travel demand,the balance of travel demand,the impact of weather on travel demand,the use and idle time of bicycles,user type and travel purpose and the single origin network(SON).The results show that the spatial and temporal distribution of users' travel demand is uneven and has obvious preferences.The exploration o users'travel pattern lays the foundation for the optimization of rebalancing services.Scheduling optimization of docking public bicycle system Combining the scanning algorithm with K-Means algorithm,the scheduling area is divided according to the workload.Regarding each public bicycle station as an inventory system,a dynamic scheduling method based on predicted inventory variation rate is proposed.According to the stationary distribution of Markov Chains,the inventory variation rate of each station during scheduling period and the inventory rate at the end of scheduling period are predicted.Mixed Integer Programming(MIP)model is established to minimize the total routing distance of repositioning vehicle.Simulated Annealing(SA)algorithm is used to solve the model for optimal scheduling scheme.Randomized experiments were conducted on public bicycle systems in Boston and Washington.The results show that,when compared to the method based on Rolling Horizon,the total routing distance at the system level can be shortened by 21.06%and 17.26%,which verifies the effectiveness of the dynamic scheduling fiamework.Scheduling optimization of dockless public bicycle system Based on the modified DBSCAN and spectral clustering algorithm,the hot spots of the system are recognized and divided.Considering time and weather factors,the GBDT algorithm is used to predict the travel demand of the system in the given period.Taking historical travel proportions into consideration,the travel demand of the system is allocated to each region.The virtua 1 stations are set up according to the travel thresholds.Referring to the docking system,the optimal scheduling scheme is generated.Taking the Mobike system in Beijing as an example,one hot spot was selected for randomized experiments.The results show that the average error rate of the forecasting model is 6.8%,and the scheduling model can shorten the total routing distance by 18.95%,which verifies the applicability of the dynamic scheduling framework.