S-Space College of Engineering/Engineering Practice School (공과대학/대학원) Dept. of Civil & Environmental Engineering (건설환경공학부) ICASP13
A multiobjective optimization based approach for RBDO
- Poirion, Fabrice; Mercier, Quentin
- Issue Date
- 13th International Conference on Applications of Statistics and Probability in Civil Engineering(ICASP13), Seoul, South Korea, May 26-30, 2019
- In this paper we present a novel algorithm in order to solve multiobjective design optimization problems of a sandwich plate when the objective functions are not smooth and when uncertainty is introduced into the material properties. The algorithm is based on the existence of a common descent vector for each sample of the random objective functions and on an extension of the stochastic gradient algorithm. It will be shown that a chance constraint optimization problem such as a RBDO problem can be written as a multiobjective optimization problem. Chance constraint optimization problems yields optimal designs for a fixed given level of probability for the constraint. However in real life problem it is not realistic to introduce a given probability because it is not known. It is more efficient to solve the problem for a whole range of probability in order to obtain an overview of the probability level appearing in the constraint effect on the solution. We show in this paper how to transform a chance constraint optimization problem into a multiobjective optimization problem and we give an illustration on simple examples.