Active redundancy allocation in complex systems by using different optimization methods

Abstract

This paper addresses the issue of optimal allocation of spare modules in complex series-parallel redundant systems in order to obtain a required reliability under cost constraints. To solve this optimization problem, an analytical method based on the Lagrange multipliers technique is first applied. Then the results are improved by using other optimization methods such as an evolutionary algorithm and an original fine tuning algorithm based on the idea of hill climbing. The numerical results highlight the advantage of combining analytical approaches with fine tuning algorithms in case of very large systems. By using such a combined technique, better solutions are obtained than those given by classic heuristic search algorithms.