Characterization of Diagnosis Algorithms for Real-time Scheduling on Distributed Processors in Support to e-Maintenance Services

This paper presents a characterization study of agent algorithms of distributed diagnosis, for real-time scheduling and on-line implementation. These parameters will be used to optimize the strategies of real-time distributed diagnosis. The characterization of execution parameters of an agent algorithm must satisfy the diagnosis time constraints related to the availability objective of the each equipment, defined in the predictive maintenance service policy. Hence, we propose to define the time characteristics and constraints, in the case of networked distributed system architecture. The diagnosis agent algorithms are then characterized for evaluating their performances, following criteria relating to the real-time execution