Abstract
This chapter presents a method for evaluating dynamic performance of multi-state systems with a general series parallel structure. The system components can be either repairable binary elements with given time-to-failure and repair time distributions, or 1-out-of-N warm standby configurations of heterogeneous binary elements characterized by different performances and time-to-failure distributions. The entire system needs to satisfy a random demand defined by a time-dependent distribution. Iterative algorithms are presented for determining performance stochastic processes of individual components. A universal generating function technique is implemented for evaluating the dynamic system performance indices. Examples are provided to demonstrate applications of the proposed methodology.