Nonequilibrium simulated-annealing algorithm applied to reliability optimization of complex systems

This paper applies an improved nonequilibrium simulated-annealing (I-NESA) technique to find: (1) the global optimum of system cost of two kinds of complex systems subject to constraints on system reliability, and (2) the optimum number of redundancies which maximize the system reliability, subject to constraints on system cost, weight, and volume in a multistage mixed system. The efficacy of I-NESA in solving both varieties of problems is demonstrated by comparing its results with those of simulated annealing (SA). I-NESA, using the Glauber algorithm and an exponential cooling schedule, provides a stable global solution to all the problems considered. The essential features of I-NESA, (1) the nonequilibrium concept while coming out of an inner iteration, and (2) incorporation of the simplex-like heuristic, make it very fast and stable in obtaining the global solution when compared to the traditional SA. Fast convergence was observed in all the problems studied. I-NESA is a useful alternative to either indirect optimization methods or to some random search techniques, in solving problems like those in this paper.