Evaluation of Collective Taxi Systems by Discrete-Event Simulation

When considering the need for an intelligent transportation mode based upon the fact that car travel is still the consumer's first choice, we decided to explore the idea of collective taxis more fully. More precisely, we are seeking to provide an autonomous high quality door-to-door service, affordable by almost everyone, covering the entire urban area simply by allocating in an optimal way more than one passenger to each vehicle and using the latest available information regarding traffic condition etc. We intend to operate intensive discrete event simulations testing multiple policies, for both centralised and decentralised management, with a fine tuning of intelligent real-time decision algorithms all together with off-line resource optimisation. Thus we shall be able to learn the system behaviour and consequently to provide an optimised structure allowing an the best behaviour for every possible demand level and geometry as well as the associated costs.