SVR响应面与进化多目标优化在结构设计中的应用

来源期刊:中南大学学报(自然科学版)2018年第7期

论文作者:胡晓兵 田昆 赵清祥 徐营利

文章页码:1650 - 1657

关键词:支持向量回归机响应面;进化多目标优化算法;正交旋转组合设计;优化系统

Key words:support vector regression response surface; evolutionary multi-objective optimization algorithm; orthogonal rotation combination design; optimization system

摘    要:针对复杂结构优化设计中常出现的低效率、低精度问题,提出支持向量回归机(SVR)响应面与进化多目标优化算法相耦合的优化系统。基于结构风险最小化推导SVR响应面建立原理,采用具有异点预测值可对比特色的正交旋转组合设计作为样本点选取法,以获得最优试验区。基于NSGA-Ⅱ范式原理建立区间偏好进化优化算法并构建优化系统框架。以45 t门机主梁为研究对象,指定5个几何参数为设计变量,对最大位移、应力进行约束,以轻量化及首阶固有频率为双目标,利用所构建的优化系统进行优化求解。研究结果表明:主梁总质量减少15.9%,首阶固有频率减少9.2%。通过不同响应面模型的效果对比、灵敏度分析及优化方案检验,验证出优化系统的高效可行性。

Abstract: Aiming at the low efficiency and low precision in complex structure optimization design, an optimization system with support vector regression(SVR) response surface coupled with evolutionary multi-objective optimization algorithm was proposed. Based on structural risk minimization, the principle of SVR response surface was deduced. Orthogonal rotation combination design with contrasting predictive value was used to select sampling points to obtain the optimal experimental area. Based on the principle of NSGA-Ⅱ paradigm, the evolutionary optimization algorithm of interval preference was established and the optimization system framework was constructed. Taking the main girder of 45 t gantry crane as the research object, five geometrical parameters were adopted as design variables, and the maximum displacement and stress were constrained. The results show that by using the optimized system, the total mass of main girder and the first-order natural frequency are decreased by 15.9% and 9.2%, respectively. The high efficiency and feasibility of the optimization system are proved by the comparison of the results of different response surface models, the sensitivity analysis and the optimization scheme validation.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号