支持向量机近似模型在船舶性能不确定度分析中的应用(英文)

A Lease Square Support Vector Machine Metamodel for Ship Performance in Uncertainty Quantification Study

棒性优化和可靠性优化中进行的不确定度分析需要大量的样本计算。当利用高精度分析工具时,其代价非常高昂。作为高精度模拟工具的近似,近似模型具有其高效性。文章基于最小二乘支持向量机(LSSVM)理论,提出了一种描述船舶性能的近似模型,分别在一维和二维不确定度分析中,验证了其精度及收敛性,并对其在不确定分析中的收敛性能进行了研究。文中结论可以为研究人员和工程师在不确定分析、鲁棒性优化和可靠性优化中提供一种新的选择。

Serving in robust design optimization （RDO） and reliability-based design optimization （RB- DO）, the uncertainty quantification （UQ） requires a large number of samples, which is very expen- sive when using high-fidelity simulation tools. As an approximation of expensive high-fidelity simu- lation codes, the metamodel has its high efficiency. This study proposes a metamodel for ship per- formance based on lease square support vector machine （LS-SVM）, validates the accuracy of the metamodel in 1D and 2D UQ studies and demonstrates the convergence of accuracy and UQ perfor- mance. The results of this study could provide researchers and engineers an option in UQ study, RDO and RBDO.