ETTO: Emergent Timetabling by Cooperative Self-organization

摘要

Cooperation is a means for multi-agent systems to function more efficiently and more adaptively. Cooperation can be viewed as a local criterion for agents to self-organize and then to perform a more adequate collective function. This paper mainly aims at showing that with only local rules based on cooperative attitude and without any global knowledge, a solution is provided by the system and local changes lead to global reorganization. This paper shows an application of cooperative behaviors to a dynamic distributed timetabling problem, ETTO, in which the constraint satisfaction is distributed among cooperative agents. This application has been prototyped and shows positive results on adaptation, robustness and efficiency of this approach.