A partition based match making algorithm for taxi sharing

Jiajian Xiao, Heiko Aydt, Michael Lees, Alois C. Knoll

Abstract: Taxi sharing can help to improve the utilisation of taxis. Passengers for sharing are always chosen with some objectives in mind, for example, to minimise the inconvenience caused due to de-tours when picking up or dropping off other passengers. In this paper, we describe a method that optimises the match making process in order to minimise the inconvenience imposed on passengers. This method is based on the idea of dividing the road network into several partitions so that a certain quality of service requirement regarding inconvenience is satisfied. More precisely, this can be considered as a constrained optimisation problem.We describe a procedure how to decide on optimal parameters for the partitioning algorithm. In addition, we analyse the theoretical maximum sharing potential of commuters in Singapore using a simulation-based approach.