Forecast and interactive balancing of a docked bike sharing system

The Bike Sharing Systems (BSS) are an integral part of the multimodal transport systems in an urban area. Those systems have many advantages such as low cost, environmental--friendly and flexibility. Nevertheless, the lack of bicycles and the lack of spaces to drop them off discourage people from using the BSS. Thus, we propose a Saturation Index ($SI$) to identify the number of bicycles for a period of time (seconds, minutes, hours) in a station; thus, based on the $SI$, the supply and demand levels are known for every station. With those levels, the number of bicycles to be moved by truck among the stations is computed by solving the transship model. Finally, a route is computed using a heuristic to minimize the travel distance of the truck among the stations. To test our approach, we used the data set of the ECOBICI system in Mexico City, and we program a computational application based on R language and Python. The results show that during the day, a station change to supply bicycles to demand them. Thus, to balance the system, the proposed approach must be run every time the managers want to balance the system.