Construction and evaluation of agent domain knowledge for flexible videoconference system

This paper discusses a method to configure agent domain knowledge base, which is used to improve flexibility of videoconference systems, where flexibility means the ability to absorb the change in the system and provide an easy-to-use service. For the knowledge (INTER Knowledge) for interagent cooperation that enables interservice adjustment of flexible videoconference systems, we propose a new architecture, Meta-INTER Model, whose aim is to dynamically switch strategies and flexibly cope with problems taking into account properties of problems and progress of problem solving. Also, we apply the Meta-INTER Model to a flexible videoconference system. We show that more advanced flexibility can be implemented by this model so that interagent cooperation in flexible videoconference systems can be executed effectively, and systems can cope with stricter constraints, compared with prior models.