An extended Birnbaum importance-based two-stage heuristic for component assignment problems under uncertainty

Abstract

The component assignment problem (CAP) is to find the optimal assignment of n components to n positions in a system that maximizes the system reliability. Due to the insufficiency and the inaccuracy of related data and system complexities, there are different types of uncertainties in real-world engineering system analysis. This paper proposes an extended Birnbaum importance (BI)-based two-stage (EBITS) heuristic using evidential network and interval-valued BI measures for solving CAPs under parametric and model uncertainties. The parametric uncertainty is the uncertainty about component reliabilities and the model uncertainty is the uncertainty about the system structure. The proposed method is applied to find the optimal assignment of valves in a ship’s fuel service system.