Diagnosability in concurrent probabilistic systems
Xiaowei Huang · Adaptive Agents and Multi-Agents Systems · 2013
Diagnosability is a key attribute of systems to enable the detection of failure events by partial observations. This paper addresses the diagnosability in concurrent probabilistic systems. Four different notions (L-, P-, A-, and AA-diagnosability) are characterised by formulas of a logic of knowledge, time and probability. Also, we investigate the computational complexities of verifying them: the L-diagnosability is NL-complete, the A-diagnosability is PTIME-complete, and the P-diagnosability is in PSPACE.