基于支持向量机和粒子群算法的稳健优化

Robust Optimization Based on Support Vector Machine and Particle Swarm Algorithm

针对实际工程中不确定性因素与产品质量特性之间不具有显式函数关系的稳健优化问题时,代理模型的精度成为关键。本文提出一种基于支持向量机代理模型和粒子群算法的稳健优化方法,采用拉丁超立方试验设计采样布点,优化问题的目标性能函数、约束函数的均值和标准差由具有自动参数优化的支持向量机模型替代,采用粒子群优化算法对稳健优化模型进行求解。以典型的两杆结构优化为例,结果表明支持向量机代理模型的综合性能比常用的响应面、BP神经网络和Kriging模型更优越,稳健优化结果比较理想,为复杂产品的不确定性设计优化提供了一种新的思路。

Aiming at addressing the robust optimization problem with implicit performance functions, the availability of efficient and accurate meta model is crucial to the success of applications of robust optimization of eomputationally intensive simulation models. A robust optimization methodology based on support vector machine （SVM） and particle swarm algorithm （[＇SO） is presented for the problems that involve high dimensional. The methodology is combined with experimental design theories, SVM approximation model and PSO. The applicability of the algorithm is demonstrated by using a two-bar structure system as case study, in which the performances of SVM were compared with those of polynomial regression （PR）, Kriging and back-propagation neural networks （BPNN）, and the results show that the prediction accuracy of SVM model is higher than those of others metamodels, and the robust optimization is accurate and efficient. The optimization methodology is effectively utilized to achieve a potential performance improvement.