Conclusion and Perspectives

Abstract

This book proposes a methodology for multi-agent systems (MAS) development that guides the user towards an implementation-independent, conceptual model set. The CoMoMAS methodology — Conceptual Modelling of MAS — has been inspired by the CommonKADS approach, which addressed the development of knowledge-based systems rather than the development of multi-agent systems (MAS). CoMoMAS can be seen as an extension of CommonKADS towards MAS development. In addition to CommonKADS run-time aspects it takes into consideration, in particular, the structuring of problem-solving knowledge guarantees flexibility. A knowledge engineering environment has been designed to apply our conceptual approach for MAS development. It integrates the commercial KADSTOOL environment and produces descriptions in an extended CML specification. A simulation environment has been realised to validate multi-agent architectures. This environment is an extension of the Michigan Intelligent Cooperation Experiment (MICE): new features are methods for the evaluation of experiments, an improved user interface and an object-oriented implementation. A case study has presented the reverse-engineering of two existing multi-agent architectures for mobile robot control.

We can summarize the empirical evidence: “Knowledge is Power”. A system exhibits intelligent understanding and action at a high level of competence primarily because of the specific knowledge that it can bring to bear: the concepts, facts, representations, methods, models, metaphors, and heuristics about its domain of endeavor. —D.B. Lenat and E.A. Feigenbaum (1987) The knowledge as power hypothesis