This article investigates diagnosability and $T$ -diagnosability for discrete-event systems modeled by unambiguous max-plus automata (UMPAs). More precisely, diagnosability requires that the occurrence of any fault can be detected within… Click to show full abstract
This article investigates diagnosability and $T$ -diagnosability for discrete-event systems modeled by unambiguous max-plus automata (UMPAs). More precisely, diagnosability requires that the occurrence of any fault can be detected within a finite number of events after the fault has occurred. $T$ -diagnosability requires that the occurrence of any fault can be detected within a delay of at most $T$ time units after its occurrence. First, we propose a polynomial-time algorithm based on the construction of a nondeterministic finite automaton over a weighted alphabet for diagnosability verification of a UMPA. Second, we prove that $T$ -diagnosability of a UMPA can be studied by reducing it to the problem of diagnosability. Third, we introduce an approach to calculate the upper on the time needed for detecting fault occurrence for a diagnosable UMPA, and its complexity is of sixth order in the number of states of the UMPA.
