Intelligent Agent Software Development Using AUML and the Descartes Specification Language

Developing software systems that are complicated and critical gives challenges to software engineers and the software industry. Intelligent software agent design is one such application domain that imposes challenges. The application of formal methods in software engineering, specifically with executable software specification languages, can involve much learning among the users of the system. Although executable software specifications tend to be advantageous over the traditional software specification process, the passive growth inside industry remains open for research efforts to come up with the hidden myths. The application of formal methods during system development of intelligent software agents makes the design of the product reliable. This paper describes an approach that combines an extension to the widely used Unified Modeling Language (UML) for design with an executable specification language. Also, the paper describes an approach that transforms agent oriented specifications written using the extended Descartes specification language into a high level design representation using Agent UML (AUML).