Assessment of Testability on Multiagent Systems Developed with Organizational Model Moise

Abstract

Trying to guarantee some level of correctness for multiagent systems (MAS) is a hard task once we are dealing with systems that inherently has properties such as autonomy, reactivity, pro-activity, and social skills. The social dimension of a MAS can be described with an organizational model, where agents are structured using concepts like groups, roles and must obey some norms. It would be useful to dimension how many tests are necessary to validate an organizational model as a metric for validation and verification processes. In this paper, we propose a method to assess the testability of an organizational model specified with Moise. Using a graphical process based on Coloured Petri Nets (CPN), the method indicates the number of test cases necessary to validate a Moise model. In a broad view, the obtained results are an important metric to evaluate the complexity of validation of a social dimension of a MAS.