A new approach for determining the most important system components and the budget-constrained system reliability improvement

Abstract

Importance measures are used for indexing system components due to their impact on the system's overall reliability. In order to identify the specific number of the most critical components, first-ranked components are singled out as the most important ones. However, importance measures consider only the influence of individual components and they are not applicable to combinations or groups of components. This common feature of importance measures is referred to in literature as one of still open issues. This paper proposes a new approach for determining the most important system components, where a whole set of components are determined simultaneously taking into account their interdependence. In systems with a large number of interdependent components, the number of the most important components which should be prevented is often limited due to the available budget. Using pre-known minimal cut sets, a mathematical model based on the Budgeted Maximum Coverage Problem is proposed. By its optimization, the simultaneous determination of all of the most important components whose total expenses do not exceed the limited overall budget is achieved. The new approach was tested by a series of experiments conducted over a set of test examples. The results of the experiments were compared with the results obtained using two commonly used cost importance measures.