Station-based, free-float, or hybrid: An operating mode analysis of a bike-sharing system

Abstract

The profit-oriented bike-sharing industry, such as Mobike in China and Limebike in the USA, has widely adopted the free-float mode, which allows users to rent and return bikes without any restrictions with regard to stations. Compared to the traditional station-based mode, this new mode improves convenience and satisfies a variety of travel demands, but it also results in higher operating costs and lower number of available bikes (the number of bikes can be found). So far, no theoretical study has been done comparing the performance levels of the two different operating modes. Some bike-sharing firms are now attempting to design a hybrid operating mode that combines the advantages of the two modes. We design a mixed-integer program to model these three operating modes under different operating conditions. The model captures some of the new features of free-float bike-sharing systems. These new features include discounts for returning bikes to designated stations, different numbers of available bikes at stations and free-float areas, differences in station-based return levels as well as in convenience levels in the three operating modes. This study also considers the interactions between the free-float and station-based modes in meeting the demand and in managing relocation processes. Extensive numerical experiments are conducted and the results reveal that in more than half of the total cases under different parameter settings, the profits of the hybrid mode outperform the profits of the other two modes. The station-based mode is suitable for settings with high convenience levels, low number of available free-float bikes, and limited budgets for bike acquisitions, whereas the free-float mode can ensure high profits with low convenience levels and a large number of bikes being available. Our models can serve as a decision-making tool for bike-sharing firms in their selection of an optimal operating mode.