Multistage Negotiation for Distributed Scheduling for Resource-Bounded Agents

Abdel-Illah Mouaddib

Progressive reasoning was introduced as another way of flexible computation to handle transient overload (resource-bounded) of real-time systems. In a system based on this technique, each time constrained agent is designed in such a way that it can adapt the quality of its solution to the available time of the whole system. The monitoring complements this system with a mechanism that determines for each agent its run-time. We address in this paper how to use Distributed Artificial Intelligence (DAI) method s for controlling self-interested and time-constrained agents where a central controller do not exist. Distributed scheduling consists of each agent locally allocates its time to build its local schedule and then it coordinates its local schedule with those of the other agents. We propose Partial Global Flexible Scheduling (PGFS) which consists of cooperation allowing a new agent to construct its local schedule and negotiation allowing the constructed local schedule to be coordinated with local schedules of the other agents. This approach is an important application of the contract-net technique to the distributed scheduling of resource-bounded and self-interested agents.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.