<![CDATA[Safety and economic factors are the two main consideration in designing a structure. The structural engineer always try to find the optimal structure design with minimum cost that satisfy the safety requirement. This safety requirement can be expressed as structural reliability that associated to a certain failure probability threshold. An integrated Reliability-based Design Optimization (RBDO) framework usually employed to minimize the cost objective function subjected to the failure probability limit. Failure probability mostly computed by using a time-consuming Monte Carlo Simulation (MCS) method. This study develops two hybrid RBDO framework, SOS-ANN and PSO-ANN, which combine the metaheuristic method, Symbiotic Organisms Search (SOS) and Particle Swarm Optimization (PSO) with a machine learning method, Artificial Neural Network (ANN). The SOS and PSO method are used to solve the discrete optimization problem. The ANN method is adopted to replace the MCS method in predicting the reliability of every solution using binary classification. A practical RBDO case of steel truss structure is used to demonstrate the performance of both SOS-ANN and PSO-ANN method in finding the optimal structural design. The results show that the SOS-ANN method outperforms the PSO-ANN method in terms of solution quality, computational efficiency and consistency.]]>
Copyrights © 2022
