Resource sharing in multi-agent systems through multi-stage negotiation

In this paper we try to solve the Temporal Resource Reallocation Problem (TRR-P) in multi-agent systems by having the agents negotiate periods of time during which they can have use of resources. Our work is based on and extends a previous work in which a multistage negotiation framework that defines the way agents negotiate over resources and time is defined. The approach in this framework suggests that the negotiating agents use a sequence of stages, each characterized by an increased chance of success and amount of exchanged information. In the same work two negotiation stages are defined each of which solves a specific range of TRR-P instances. In this paper we propose a third negotiation stage which is more sophisticated than the first two and is able to solve a larger range of TRR-P instances.