Selection of Series System Components to Maximize Reliability

The paper tackles the problem of maximizing the reliability of a series system by an opportune choice of components. Each type of component must be selected among the available alternatives for that component whereas a fixed amount of budget can not be overcome. The problem can be formulated by a binary non linear programming model and it is equivalent to a knapsack problem with multiple-choice constraints, well known to be NP-hard. An exact algorithm is proposed for solving large dimension problems to the optimum in a short time. The algorithm efficiency is finally compared with the recent heuristics proposed in literature to approach the same problem