A stochastic programming model for free-floating shared bike redistribution considering bike gathering

We study a stochastic free-floating shared bike redistribution problem, which first gathers bikes randomly scattered around the hotspots of a city and then transships the gathered bikes among the hotspots to achieve a better supply–demand balance. In the literature, the bike gathering operation and the associated stochasticity, although significant for shared bike redistribution practice, are rarely considered. To fill the gap, we propose a two-stage stochastic programming model that explicitly incorporates the bike-gathering cost with two formulations. In the planning stage, the deployment of gathering sites and the allocation of gathering staff are determined. In the operation stage, the gathering, transshipment and replenishment of shared bikes are optimized based on the decisions of the planning stage and the realization of the uncertain supply and demand of free-floating shared bikes in space. With a case study based on real data from the Mobike company in Beijing, China, we illustrate the real-world applications of our proposed model, highlight the importance of considering the cost of bike gathering when planning free-floating shared bike redistribution, and gain managerial insights for better field practices.