Integrating Multiple Reasoning in Intelligent Agent Systems

Article type: Research Article

Authors: Kwok, A.D. | Norrie, D.H.^{; *}

Affiliations: Division of Manufacturing Engineering, Faculty of Engineering, The University of Calgary, Calgary, Alberta, Canada T2N 1N4

Correspondence: [*] To whom correspondence should be addressed.

Abstract: Intelligent agent systems have the potential for integrating diverse reasoning approaches and emerging computer technologies to solve the complex problems of coordination and control in large distributed systems. The Intelligent Agent Object (IAO) System has been developed to provide a generic framework for integrating multiple reasoning and multiple programming paradigms in intelligent agent applications. This development environment currently has forward, backward, or hybrid chaining of production rules, with certainty values, as well as fuzzy logic, available for the intelligent agents created. Additional reasoning approaches are to be subsequently added. Inferencing is carried out without a conventional inference engine through a new technique of message inferencing that dynamically interlinks the four inferencing functions: variable interpretation; rule chaining; execution tracing; object state updating. Both rule-based programming and procedural programming are integrated within a uniform object-oriented structure. Propagation-, access-, and group-oriented programming are also implemented. An intelligent agent-based prototype of an automated guided vehicle operational control system has been developed and tested to demonstrate the capabilities of the IAO development system and further explore the potential for intelligent agent systems.

DOI: 10.3233/ICA-1993-1106

Journal: Integrated Computer-Aided Engineering, vol. 1, no. 1, pp. 83-90, 1993