Collaborative Optimization Design Considering variability in design variables

• Main Authors: Po-Chang Lu, 盧柏璋
• Other Authors: 廖國偉
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/63967686240906968572

碩士 === 淡江大學 === 土木工程學系碩士班 === 97 === This paper presents a new approach to evaluate the Reliability-Based Design Optimization (RBDO) via multi-level optimization process. RBDO incorporates probabilistic analysis into optimization process, providing an optimum design with a better chance of staying in the feasible design space. One of the biggest drawbacks of applying RBDO to practical problem is its high computational cost that is often impractical to industries. Many approaches have been proposed to overcome this issue such as the single-loop approach and the response surface method. This study utilizes the concept of Multidisciplinary Design Optimization (MDO) and extends it to the problem of RBDO. The proposed RBDO is two-folds: the first step is to decompose reliability analysis and optimization; the second step is to form a coordinating center as a controlling unit to monitor above analyses in parallel. This is a big improvement in regard of computation efficiency because optimization and reliability analysis is required to be conducted sequentially in the single-loop algorithm. The accuracy of the proposed method is verified in the numerical examples.