Probabilistic approach to performance evaluation of a series-parallel system

The paper deals with modeling and performance evaluation of a series-parallel system using Markov Birth-Death process and probabilistic approach. The system consists of four subsystems arranged in series-parallel with three possible states, working with full capacity, reduced capacity and failed. Through the transition diagrams, systems of differential equations are developed and solved recursively via probabilistic approach. Explicit expressions for steady-state availability are de-rived. Availability matrices for each subsystem have been developed to provide various performance values for different combinations of failure and repair rates of all subsystems. The results from this paper will enhance the system performance and useful for timely execution of proper maintenance improvement, decision, planning and optimization.