Reliability modelling and simulation of switched linear system control using temporal databases

Abstract

In reliability modelling with redundancies, it is assumed that the system would have sets of failed, standby, and working subsystems at any time. This study deals with the case of stochastic switched linear system comprising series-parallel interconnected subsystems for which operational safety conditions are specified. The subsystems are assumed to admit the same control inputs and undergo the same state transitions. Thus, constructing a subsystem Markov model and matching its parameters with the specified safety factors provides the basis for the entire system analysis. For the system simulation, temporal databases and predictive control algorithm are designed. The simulation results are analyzed to assess the reliability of the system behaviour. Graphs of the system behaviour indicate cases of highly oscillatory and fairly stable trajectories from which optimal-time controls are deduced. The study demonstrates use of temporal databases in practical analysis of such stochastic system, highlighting the dynamics beyond results of theoretical analysis. The study has applications, which include optimal design of fault-tolerant real-time switching systems control and modelling embedded micro-schedulers for complex systems maintenance.

Keywords: Switched linear system, Markov model, predictive control, reliability, safety factors

Journal of the Nigerian Association of Mathematical Physics, Volume 19 (November, 2011), pp 69 – 76