Matchmaking optimisation for car sharing services

Abstract

With the continuous growth of our cities it became obvious that personal car ownership is not sustainable on the long run. Personal cars are only used a couple of hours per day, the rest of the time occupying a parking slot, either on the street or in an underground garage. It is not surprising thus that car sharing services became more and more popular in the last few years, decreasing the number of vehicles needed to support the mobility of city inhabitants. In traditional car sharing, clients select manually a vehicle from a pool of available cars shown on a map. However, this will not result in a globally optimal car distribution. In this paper we present an ILP-based solution to optimize the matchmaking between cars and clients, so as to minimize the maximum or the average distance clients have to walk to reach their allocated vehicle. The presented model can be later enhanced to help car sharing service providers to tackle the complex problem of optimal fleet sizing.