Dynamic Rebalancing Optimization for Bike-sharing Systems: A Modeling Framework and Empirical Comparison

Bike-sharing systems have been implemented in multiple major cities, offering a low-cost and environmentally friendly transportation alternative to vehicles. Due to the stochastic nature of customer trips, stations are often unbalanced, resulting in unsatisfied demand. As a remedy, operators employ trucks to rebalance bikes among unbalanced stations. Given the complexity of the dynamic rebalancing planning, this topic has received significant attention from the Operations Research community. However, the planning problem requires significant simplifications such that optimization models remain computationally tractable. As a result, existing models have used a large variety of modeling assumptions and techniques regarding decision variables and constraints. Unfortunately, the impact of such assumptions on the solutions’ performance in practice remains generally unexplored. Indeed, existing simulation models to evaluate the performance of planning strategies also rely on simplifications, such as the aggregation of trips within the same time-period, therefore ignoring the original chronological sequence of trip demand.

In this paper, we first systematically survey the literature on rebalancing problems and their modeling assumptions. We then propose a general modeling framework for multi-period dynamic rebalancing problems that can be easily adapted to different assumptions, including trip modeling, time discretization, trip distribution, and event sequences. We develop an instance generator to synthesize realistic station networks and customer trips, as well as a more realistic simulator to evaluate the operational performance of rebalancing strategies. Finally, extensive numerical experiments are carried out to analyze the effectiveness of various modeling assumptions and techniques. In this way, we identify the assumptions that empirically provide the most effective rebalancing strategies in practice. In particular, a set of specific trip distribution constraints as well as event sequences ignored in the previous literature seem to provide particularly good results.



View Dynamic Rebalancing Optimization for Bike-sharing Systems: A Modeling Framework and Empirical Comparison