Transforming Analysis Models into Design Models for the Multiagent Engineering Systems (MASE) Methodology

Agent technology has received much attention in the last few years because of the advantages that multiagent systems have in complex, distributed environments. For multiagent systems are to be effective, they must be reliable, robust, and secure. AFIT\u27s Agent Research Group has developed a complete lifecycle methodology, called Multiagent Systems Engineering (MaSE), for analyzing, designing, and developing heterogeneous multiagent systems. However, developing multiagent systems is a complicated process, and there is no guarantee that the resulting system meets the initial requirements and will operate reliably with the desired behavior. The purpose of this research was to develop a semi-automated formal transformation system for the MaSE methodology, as one part of formal agent synthesis, that derives the system design based on the analysis. Since each transform in the transformation system preserves correctness, the designer can be sure that the resulting system design is correct with respect to the system specification. A secondary goal of this research was to develop a proof-of-concept module for agentTool that implements the transforms