A cellular evolutionary approach applied to reliability optimization of complex systems

This paper proposes an innovative approach using cellular evolutionary strategies (CES) to solve three types of reliability optimization problems: redundancy (number of redundant components), component reliability, and both redundancy and component reliability. In general, these problems are formulated as mixed-integer nonlinear programming problems with one or several constraints. CES combine evolution strategy techniques with concepts from cellular automata (CA) to solve optimization problems. CES were designed to find the global optimum or "near" optimum for complex multi-modal functions where traditional optimization techniques have shown poor performances, or simply have failed. The new technique has been applied to several typical problems with results better than previously reported and very close to the optimum solution.