A stochastic simulation model for the optimization of the taxi management system

Abstract

The paper presents a simplified simulation model of the operation of a taxi system. The model retains the main features of a real taxi transportation system and despite its simplicity examines the system behavior in different conditions. It was shown that for every request generation rate a critical number of taxis in disposal could be determined. If the real number of taxis is lower than the critical number, the queue of pending requests grows in an unlimited way. On the other hand, if the real number of taxis is significantly higher, the service level is clearly not better and leads to the waste of taxi drivers' time and fuel. The presented model can be regarded as a queue system; therefore, the well-known queue theory is used to describe its nature. However, this approach has some practical limitations coming from incomplete knowledge on real transportation demands, which additionally undergo significant fluctuations. A method, which optimizes the assignment of vacant taxis to the pending requests was also introduced. It was proven that this method mitigated the influence of the above-mentioned limitations.