A multiagent knowledge-based recommender approach with truth maintenance

Abstract

This thesis investigates a way of using knowledge in dynamic and distributed domains for supporting recommendation, keeping the consistence of the decision knowledge that change over time. We propose the use of a multiagent knowledge-based recommender approach capable of dealing with distributed expert knowledge in order to support travel agents in recommending tourism packages. Agents work as experts cooperating and communicating to each other in the recommendation process. Each agent has a truth maintenance system (TMS) component that helps the agents to keep the integrity of their knowledge bases.