Coordination of Temporal Plans for the Reactive and Proactive Goals
Résumé
In the collaborative multi-agent systems where agents' goals are prioritized, an agent often needs to be more flexible in the pursuit of its own goals, for the sake of the higher priority goals of the others. In such situations, finding an optimal plan under the constraints imposed by the other agents' plans, is really an important area which needs more investigation. This work aims to fill this gap by studying two different situations in a similar context. In the first situation, which we call Coordinated Planning, an agent has to compute a temporal plan for the achievement of its own goals, but without violating the constraints of another agent's plan, and utilizing where possible the cooperative opportunities offered by the latter. For this purpose, we propose Coordinated-Sapa an extension to the well known planner Sapa, which aims to deal with negative and positive interactions. In the second situation, which we call Proactive-Reactive Coordination, an agent has to modify its plan in order to remove any conflicts with the plan of another agent, having high priority goals. We propose a plan merging algorithm supported by a sound plan repair technique to solve this problem.