Let’s shuffle: Facility Optimal Location for Stations within Bicycle Sharing Systems in the City of Buenos Aires after the pandemic

Abstract

People’s habits have changed after the pandemic and cycling around the city of Buenos Aires is no exception. This thesis leverages literature on Capacitated Facility Location Problems (CFLP) to build an optimal bike-sharing network to minimize the total system’s cost. The objective is to decide which stations should be left open to meet projected demand in the worst-possible cases, ensuring that users do not have to walk more than a predefined distance to the facility that is closest to them. Results suggest that there is an excess of stations in the downtown area and idle capacity that could be relocated in peripheral areas, reflected by a positive load factor increase of 2x after the optimization is done. The solution shows that up to 70% of total costs could be saved after using our optimization model, by closing down facilities while meeting demand. While total cost is estimated as the budget that needs to be invested to ramp up the system from scratch, it is a useful metric that shows us how the network could be optimized taking away stations from overcrowded areas without losing any of the current demand. All of these bike-sharing facilities could be relocated to areas that have a low-density of bikes, improving access to the cycling system in the city of Buenos Aires.