Agent Negotiation Strategies for Composing Service Workflows

Composing services into executable workflows presents many challenges. One of the most difficult challenges is deciding which option among several available workflows is the best value for the user. Cost, reliability, user preferences, and other factors must be balanced to develop the best recommendation. In this paper we describe a strategy for developing optimal recommendations in a system where each service is represented by an intelligent agent that employs a bidding strategy that balances its own interest and its relationship to other agents in the proposed workflow. This offers opportunities for agents to not only compete with each other, but also to form informal ad-hoc coalitions that cooperate to win more work.