AI Communications 30 (2017) 37–52 37 DOI 10.3233/AIC-170722 IOS Press Coordinating open ﬂeets. A taxi assignment example Holger Billhardt a,∗ , Alberto Fernandez a , Marin Lujak b , Sascha Ossowski a , Vicente Julian c , Juan F. De Paz d and Josefa Z. Hernandez e a CETINIA, University Rey Juan Carlos, Madrid, Spain E-mails: holger.billhardt@urjc.es, alberto.fernandez@urjc.es, sascha.ossowski@urjc.es b IMT Lille Douai, Douai, France E-mail: marin.lujak@imt-lille-douai.fr c DSIC, Universidad Politecnica de Valencia, Valencia, Spain E-mail: vinglada@dsic.upv.es d BISITE, Universidad de Salamanca, Salamanca, Spain E-mail: fcofds@usal.es e DIA, Universidad Politecnica de Madrid, Madrid, Spain E-mail: phernan@ﬁ.upm.es Abstract. Nowadays, vehicles of modern ﬂeets are endowed with advanced devices that allow the operators of a control center to have global knowledge about ﬂeet status, including existing incidents. Fleet management systems support real-time decision making at the control center so as to maximize ﬂeet performance. In this paper, setting out from our experience in dynamic coordination of ﬂeet management systems, we focus on ﬂeets that are open, dynamic and highly autonomous. Furthermore, we propose how to cope with the scalability problem as the number of vehicles grows. We present our proposed architecture for open ﬂeet management systems and use the case of taxi services as example of our approach. We carried out some experiments, which showed our proposed algorithm outperform the most common assignment method both in waiting times of clients and taxi costs. Keywords: Multiagent systems, coordination, open systems, dynamic ﬂeet management, dynamic optimization 1. Introduction The increase of human mobility and freight trans- portation in urban environments presents one of the challenges major urban cities in Europe and all over the world are faced with in today’s society. It is one of the causes of congestion problems, inefﬁciencies in logis- tics and energy use, and air pollution in modern cities [8,24]. To approach this challenge, innovative trans- portation solutions are required that allow for a more efﬁcient use of resources (vehicles, energy resources, roads, etc.) but that assure at the same time ﬂexible mo- bility solutions for both citizens and freight distribu- tion. The idea of smart cities presents new challenges and requires new solutions related to trafﬁc and trans- port. As a direct result, in the last years more and more systems that promote the shared use of vehicles have * Corresponding author. E-mail: holger.billhardt@urjc.es. begun to emerge [17]. Solutions like public bicycle ser- vices, bike or car sharing systems, or applications like UBER, providing taxi services through “free” drivers, have the objective to improve human mobility and at the same time reducing its cost. Also in the domain of freight distribution in the business sector the idea of “ﬂexible” ﬂeets that are composed on the ﬂy by vehi- cles from possibly different owners has emerged. The goal is again to optimize the use of available resources but also to increase the ﬂexibility in providing services with more and more demand ﬂuctuations. We call this type of solutions open ﬂeets. Similar to the traditional ﬂeet concept, an open ﬂeet is operated by some entity that manages and coordinates the use of a limited set of resources in order to provide a spe- ciﬁc transportation service. However, open ﬂeets ex- tend the traditional ﬂeet concept towards a new dimen- sion of openness: vehicles may join and leave the ﬂeet at any time, and the capacity of the operator to control the ﬂeet in its entirety may vary considerably. Both of 0921-7126/17/$35.00 © 2017 – IOS Press and the authors. All rights reserved