Reliability-based MOGA design optimization using probabilistic response surface method and Bayesian neural network

In this paper, an effective optimization approach, which integrated the probabilistic surrogate model, non-dominated sorting genetic algorithm, and reliability index method, is proposed to multi-objective reliability-based design optimization. To reduce the computational cost and improve the efficiency of the optimization process, the problem can be surrogated by probabilistic response surface method and Bayesian neural network as high fidelity metamodel with statistical modelling method. After verification through the simulation results on numerical test problem, these techniques have been applied to engineering problem in optimizing simultaneously multi-performances or objective functions subject to reliability constraints.