Metrics for evaluating agent oriented software engineering model

Agent oriented software engineering (AOSE) is an efficient methodology for open, complex and distributed system in networking, academic, medical, industrial and commercial areas. In order to implement AOSE methodology for different problem domains, the developing system architecture has to be considered as a multi-agent system (MAS). Several researchers have been agreed that in order to improve the performance of AOSE development methodologies, the corresponding MAS should be well designed, described and configured. However, efficient MAS metrics of its different attributes is the best way to measure the performance and quality of MAS architecture. Therefore, in this paper we have tried to define a suit of metrics to measure certain quality attributes, i.e., safety or trust, openness and efficiency of MAS architecture. Proposed metrics are evaluated in different problem domain and system architectures.