Abstract
Issue Title: Special Issue: Innovations in Applied Artificial Intelligence Meeting scheduling (MS) represents an important real-world group decision application that denotes one of the actual combinatorial problems. Solving this problem consists of scheduling all the meetings while satisfying all the constraints related to both the users and the meetings. However, given human nature, the solution is usually delineated by the encountering of conflicting preferences. Most of existing research efforts allow the relaxation of the users' preferences in order to reach an agreement between all the participants, which is not always possible. In addition, they do not deal with the achievement of any level of local consistency to enhance the efficiency of the solving process, and finally, they do not address the real difficulty of distributed systems, which is the complexity of message passing operations. Here we propose a new approach to facilitate and streamline the scheduling meetings process in any organization. This approach is based on the distributed reinforcement of arc consistency model, which takes into account the difficulties mentioned above. The present work focuses mainly on satisfying meetings hosts' preferences as much as possible, while taking into consideration all users' availability. The underlying selfish protocol is able to efficiently reach the best solution for the host of the meeting (according to the predefined criteria) whenever possible. This process is achieved with the minimal number of exchanged messages and while retaining as much of the privacy of the involved users as possible. An experimental comparative analysis divulges that our approach is scalable and worthwhile especially for strong constraints.[PUBLICATION ABSTRACT]