Optimization Of Layout And Configuration For Public Bike Rental Station

In recent years, with the influence of the fog-haze weather becoming increasingly serious, the concept of "green transportation" has attracted more and more attention. Because of the advantages of flexibility arrangement, good for health,relieving the traffic congestion, the public bicycle traffic is an important part of green transportation. The optimization of layout and allocation of rental stations is one of the most important problems in the construction of public bike system, especially the quadratic optimization of the established public bicycle system in some cities.Reasonable layout and configuration of the rental stations will directly contribute to solving the problems such as "execution hardship in renting or returning the bike".The main researches in this thesis are as follows:1. A location strategy based on the qualitative and quantitative methods at aspects of fuzzy screening and costs is proposed: firstly filter, then optimization. It means that qualified locations meeting the qualitative requirements were selected firstly based on the factors related to the surrounding environment of the stations, then the final stations are determined by an optimization model of minimization of the travel cost and construction cost.2. An optimization model of bicycle configuration based on the balance of bicycles is put forward through analysis of the operational data of the public bike systems which have the characteristics of the fluctuation and imbalance between the different stations, as well as different time periods. It sets up an optimization model which aims to achieve the least cost of bike purchase and system operation. The model proposes to replace the cost of vehicle scheduling with reword ,which encourages users to take an active part in the process of scheduling.3. The above models are validated on reducing the cost of users' trip and increasing the utilization rate of each bike with some simulation experiments by comparing with the results that are got without adopting an optimization. The results show that the models and methods presented in this thesis are feasible and effective.