Multi-agent approach for on-demand transport problem in a city

Abstract

The traffic conditions in cities vary according to time and this is reflected on the travel speed that changes on certain roads during the day. A system, treating a real on-demand transport (ODT)problem, must take into account the historical traffic data and the stochastic traffic data disrupting the usual traffic situation. Hence, obtaining a system able to estimate the travel speed at a given time and to provide a solution based on reliable information. Thereby, the passengers are serviced better. In this article, we propose a multi-agent approach to resolve the ODT problem in a city where the travel speed on some roads is time-dependent. In this approach, the travel speed at a specific time is estimated depending on the historical traffic data of the city. This is in order that our ODT system provides the most possible reliable results. The experimental results confirm the necessity to take into account the time-dependent travel speed in order to increase the clients satisfaction.