Negotiation Strategies in Multi-Agent Systems for Meeting Scheduling

An agent not always can single-handed finish a certain task, sometimes he has to interact with other agents to achieve the desired outcome. This interaction can be a time request, in which the agent must spend with the other agent. To address this task, this research focuses on using a multi-agent system for automatic scheduling. During this task, agents may not agree upon the date of a meeting or take too much time to decide what action to execute. In order to solve these conflicts, this study applies a counterproposal technique and representation of user preferences to implement two different negotiation strategies in a multi-agent system for an automatic meeting scheduler. The scenarios of this work indicate that combining different strategies at certain moments can increase the chance of success during commitment definition, reduce the amount of information shared among agents, avoid privacy issues, raise the expressiveness of agents acceptance region and remove the necessity of a strategy selector.