Security Testing for Multi-Agent Systems.

A combination of formal methods and security testing for communication layer of multi-agent systems is proposed and studied. We start with security property called bisimulation process opacity. Unfortunately, this property is undecidable in general so we propose its more realistic variant based on simulation and tests and testing. A test represents an attacker’s (i.e. possibly one of the agents) scenario to obtain some confidential information on systems. Here we consider system to be secure if it cannot be compromised by a given test or set of tests. By test we can express also capabilities of an attacker related to time properties such as time measurement accuracy, duration of tests (attacks), complete lack of attacker’s time information and so on. At the end we state a decidability result for testing.