Conditional Value at Riskbased Multidisciplinary Robust Design Optimization

On account of the coupling mechanism of the complex system, it is intractable to effectively solve multidisciplinary robust design optimization (MRDO) under uncertainty. Conditional Value at Risk (CVaR) is a familiar tool for robust optimization in financial engineering and it provides an alternative plan for traditional robust methods. This paper develops an efficient framework for MRDO by combining CVaR method. The collaboration model is employed to filter the samples which satisfy the coupled state equations. The filtered samples are taken to establish the objective function and constraint function which involve coupled state parameters. A gradient-based Monte Carlo simulation (MCS) method is employed to drive the optimization process. The details of the CVaR-MRDO optimization framework are elaborated. A mathematical example and a liquid cooling battery thermal management system (BTMS) are used to demonstrate the applicability and effectiveness of the proposed method.