Opportunities for Learning in Multi-Agent Meeting Scheduling

Elisabeth Crawford and Manuela Veloso

In this paper we explore opportunities for learning in Multi-Agent Meeting Scheduling. We view this multi-agent task as fully distributed with several challenging characteristics: (i) agents have ownership of their own calendars; (ii) agents exchange information among each other with the goal of finding an open meeting time; (iii) agents can negotiate multiple meetings concurrently. We have implemented a negotiation strategy in which the agents communicate all their available time slots. This "open negotiator" is designed to reflect the open calendar approach of Microsoft Outlook in a distributed setup. We show where this negotiation strategy fails to lead to efficient scheduling and discuss how agents could use learned information to improve social welfare and scheduling performance.

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.