Reliability optimization design using a hybridized genetic algorithm with a neural-network technique

In this paper, we examine an optimal reliability assignment/redundant alloc ation problem formulated as a non-linear mixed integer programming (nMIP) m odel which should simultaneously determine continuous and discrete decision variables. This problem is more difficult than the redundant allocation pr oblem represented by a nonlinear integer problem (nIP). Recently, several r esearchers have obtained acceptable and satisfactory results by using genet ic algorithms (GAs) to solve optimal reliability assignment/redundant alloc ation problems. For large-scale problems, however, the GA has to enumerate a vast number of feasible solutions due to the broad continuous search spac e. To overcome this difficulty, we propose a hybridized GA combined with a neural-network technique (NN-hGA) which is suitable for approximating optim al continuous solutions. Combining a GA with the NN technique makes it easi er for the GA to solve an optimal reliability assignment/redundant allocati on problem by bounding the broad continuous search space by the NN techniqu e. In addition, the NN-hGA leads to optimal robustness and steadiness and d oes not affect the various initial conditions of the problems. Numerical ex periments and comparisons with previous results demonstrate the efficiency of our proposed method.