An algorithmic approach to increased reliability through standby redundancy

The authors address the issue of optimal redundancy allocation, viz, improving the reliability of a system by adding cold-standby spares, through an algorithmic approach that uses the unique characteristics of an assumed underlying failure distribution. Although the approach applies to systems with mixed parallel and series configurations, the discussion is limited to a series system of components. The lifetime distributions of individual parts are of the phase type. This class of probability distributions is chosen for its ease of numerical implementation. The formulation of the reliability enhancement problem is applied, and several heuristic design algorithms are examined. >