Reliability-redundancy optimization using simulated annealing algorithms

Purpose - This paper aims to present a simulated annealing (SA) algorithm to search the optimal solution of reliability-redundancy allocation problems (RRAP) with nonlinear resource constraints. Design/methodology/approach - The developed SA algorithm is coded in C++ and is applied to reliability design problems which include the series system (P1(a) and P1(b)), the series-parallel system (P2), and the complex (bridge) system (P3). The numerical experiments are executed on an IBM-PC compatible with a Pentium IV 2.0 GHz. The results are compared with those of previous studies. Findings - The SA algorithm can find better solutions comparable to the previous studies in all problems except the problem P1(b). The difference on the order of 10 24 between the best and worst for all problems indicates good solution convergence of the SA algorithm. Note that the CPU times for these problems are within a few seconds by Pentium IV 2.0 GHz (P1(a) = 2.78 sec; P1(b) = 3.37 sec; P2 = 1.38 sec, and P3 = 1.40 sec). Originality/value - The application of the SA is expanded to the RRAP, which can help reliability engineers design the system reliability.