Reliability redundancy allocation: An improved realization for nonconvex nonlinear programming problems

The redundancy allocation problem, which seeks to find the optimal allocation of redundancy that maximizes system reliability, is one of the representative problems in reliability optimization. The problem can be formulated as a nonconvex integer nonlinear programming problem. This paper presents an efficient branch-and-bound approach to solve this problem, where the considered system is coherent, that is, the objective and constraint functions have monotonic increasing properties. The proposed method is based primarily on a search space elimination of disjoint sets in a solution space that does not require any relaxation of branched subproblems. The main advantage of the proposed method is flexibility (i.e., it does not rely on any assumptions of linearity, separability, single constraint, or convexity) which make the method adaptive to various applications. Numerical experiments demonstrate that the method is superior to the existing exact algorithms for redundancy allocation problems in terms of computation time.