Car allocation in a rotary car park
by
Mark Fackrell
University of Melbourne
Coauthors: Peter Taylor, Kiran Chellaram, Kohleth Chia, Ginevra Morgan

Rotary "ferris wheel" style car parks provide a space efficient way of parking cars in a crowded environment. The problem is, however, determining where the cars are to be parked as they arrive. Clearly, if the cars are simply put in adjacent parks, the car park could become unbalanced rather quickly. Engineering for this situation would be quite costly. We set up the car allocation problem as a Markov decision process in order to minimize imbalance (and to a lesser extent the amount the car park turns) and demonstrate that this is a better approach.

Date received: December 17, 2009