Robust Design in Structural Mechanics

Abstract

Many of the multi-objective optimization problems are often subject to parameters with uncertainties and noises. In such cases, to identify the robust solutions we generally add small amounts of noise and evaluate them with Monte Carlo simulation. In this paper, we suggest a new methodology for solving these types of multi-objective optimization problems. This methodology consists of increasing the objective function space with robustness functions in order to find robust and optimal solutions. The multi-objective optimization problem is solved with an evolutionary algorithm. A neural network is used to significantly reduce the computational time, in particular for the robustness function evaluations.

Keywords:

• Multi-Objective Optimization
• Neural Networks
• Genetic Algorithms
• Uncertainties
• Robustness