Collective intelligence for demand-responsive transportation systems: a self organization model

Abstract

Today's public transportation systems are subject to many requirements that are difficult to reconcile. They are deterministic and based on frequencies and routes that are set in advance. In this paper, we propose a self-organization model in order to carry out a global adaptability that is able to maintain a good dispatching of the fleet of vehicles in a purely dynamic Demand-Responsive Transportation Systems. Our proposal consists on a decentralized approach and a multi-agent system to model the environment. The behaviour of the proposed system is derived from a collective arrangement of the various individual activities, enabling vehicles agents, with respect to a given demand, to calculate the overcosts via an offer protocol and following an insertion heuristic.