Research On Borrow-Return Characteristics And Scheduling Model Of Urban Bike-Sharing System

Public transport has get great support from the national and all levels of government due to its characteristics of large capacity, high efficiency, low cost, low pollution. Developing all-round, multi-level integration of public transport is the effective measure to alleviate various traffic problems caused by the contradiction between supply and demand of urban traffic. Therefore, each region is actively exploring the development mode of public transport, including introduction and development of urban bike-sharing system. That's because, first, the public bicycle itself as a way of public transportation is an effective supplement to the traditional public transport. Second, the public bicycle plays an important role in transferring excessive passenger flow after the public transportation with large capacity, solving the problem of the last kilometer. However, the public bicycle system in domestic urban generally faces the problem of not borrowing and returning in time during the peak time at some sites, which affects the efficiency of the system and the satisfaction of users seriously. On the basis of some related theory researches, this thesis focuses on borrow-return characteristic and scheduling model of urban bike-sharing system, aiming at improving efficiency of system, reducing operating costs and improving the satisfaction of users. This research has important research significance and practical value.First, the concept of urban bike-sharing system is defined, introducing this system from its composition, function orientation and site facilities. Based on the operational data of Jiangning district in Nanjing, in order to grasp user's borrow-return features, this research analyze borrow-return times features, time characteristics and turnover characteristics after eliminating the invalid data.Secondly, this research analyzes the scheduling between urban public bicycle station, divides into types of scheduling and determines the scheduling model. With the purpose of the strongest overall stability of system and latest scheduling time during peak time, a public bicycle rental station scheduling demand model is built. In order to solve this model, genetic algorithm with decimal coding, competition selection and arithmetic crossover is designed.Thirdly, with the purpose of minimize the distance of the dispatch vehicle, taking scheduling vehicle capacity constraints and dispatch vehicles supply constraints into consideration, establishing a static scheduling vehicle routing optimization model. Based on the basic ant colony algorithm, an improved ant colony algorithm is used to solve this model.Finally, the characteristics of dynamic scheduling problem is clarified, scheduling service time window of each site in dynamic scheduling is determined, the relationship between scheduling demand and scheduling time of each site is calculated.