基于小波尺度函数的WSK-SV算法 及其气动性能预测
WSK-SV algorithm based on scaling function of wavelet and its prediction for aerodynamic performance
-
摘要: 提出了一种将小波的尺度函数与SV(support vector)算法相结合的WSK-SV(wavelet scaling kernel-support vector)新算法,并将Daubechies小波以及Shannon小波的尺度函数分别构成尺度核函数,而且分别作为SV算法中一个可容许的支持向量核函数使用.该算法充分利用了Daubechies小波函数的紧支集与正交等特点以及小波的MRA(multi-resolution analysis,多分辨分析),并注意了尺度核函数能够满足Mercer条件.该算法除了具有通常SVM(support vector machine)所具有的优点外,还具有很好的收敛性以及泛化能力,能够有效地提高学习与预测效率.典型算例选取了不同的小波尺度函数,数值计算表明:在一维、二维和三维问题中,这些小波的尺度函数均可以用于WSK-SV算法,进而显示了这个新算法的可行性与通用性.Abstract: A new wavelet scaling kernel-support vector (WSK-SV) algorithm based on scaling function of wavelet and support vector (SV) algorithm was presented in this paper,which firstly took Daubechies and Shannon wavelet-scaling kernel function for a kind of admissible support vector kernel,respectively.WSK-SV algorithm possesses such properties as compactly supported wavelet bases,orthogonal bases,multiresolution analysis (MRA),and satisfies Mercer condition for scaling kernel function.This algorithm not only has the advantages of general support vector machine (SVM),but also has a good convergence and excellent capacity of generalization,which can improve the learning efficiency and predicting ability.Numerical experiments demonstrate that this proposed WSK-SV algorithm is feasible and effective.
-
[1] CHEN Naixing.Aerothermodynamics of turbomachinery:analysis and design[M].Singapore:John Wiley & Sons,2010. [2] 王保国,刘淑艳,李翔.基于Nash-Pareto策略的两种改进算法及其应用[J].航空动力学报,2008,23(2):374-382. WANG Baoguo,LIU Shuyan,LI Xiang.Two improved algorithms based on Nash-Pareto strategy and their applications[J].Journal of Aerospace Power,2008,23(2):374-382.(in Chinese) [3] Deb K,Pratap A,Agarwal S,et al.A fast and elitist multi-objective genetic alogorithm:NSGA-Ⅱ[J].IEEE Transactions on Evolutionary Computation,2002,6(2):182-197. [4] Kim J H,Choi J H,Husain A,et al.Multi-objective optimization of a centrifugal compressor impeller through evolutionary algorithms[J].Journal of Power and Energy,2010,224(5):711-721. [5] 王保国,吴俊宏,刘淑艳.流场特性预测的两类高效方法[J].航空动力学报,2010,25(8):1763-1767. WANG Baoguo,WU Junhong,LIU Shuyan.Two effective methods for flow-field performance forecast[J].Journal of Aerospace Power,2010,25(8):1763-1767.(in Chinese) [6] Vapnik V N.The nature of statistical learning theory[M].New York:Springer,1996. [7] Suypens J A K,Gestel T V,Brobanter J D,et al.Least squares support vector machines[M].Singapore:World Scientific,2002. [8] Schölkopf B,Burges C J,Smola A J.Advances in kernel methods:support vector learning[M].Boston:MIT Press,1999. [9] Mallat S.A wavelet tour of singal processing[M].2nd ed.San Diego:Academic Press,1999. [10] WANG Baoguo,WU Junhong.Construction of Daubechies wavelet and its application in shock capturing .GyeongJu:9th International Symposium on Experimental and Computational Aerothermodynamics of Internal Flows (ISAIF),2009-3D-1,2009. [11] WU Junhong,WANG Baoguo.A new numerical method based on wavelet detection and its application .GyeongJu:9th International Symposium on Experimental and Computational Aerothermodynamics of Internal Flows (ISAIF),2009-2D-3,2009. [12] 王保国,吴俊宏,朱俊强.基于小波奇异分析的流场计算方法及应用[J].航空动力学报,2010,25(12):2728-2747. WANG Baoguo,WU Junhong,ZHU Junqiang.Method based on wavelet singularity analysis for complicated flow and its application[J].Journal of Aerospace Power,2010,25(12):2728-2747.(in Chinese) [13] 王保国,刘淑艳,钱耕.一种小波神经网络与遗传算法结合的优化方法[J].航空动力学报,2008,23(11):1953-1960. WANG Baoguo,LIU Shuyan,QIAN Geng.Optimitation method by combination of wavelet neural networks and genetic algorithm[J].Journal of Aerospaces Power,2008,23(11):1953-1960.(in Chinese)
点击查看大图
计量
- 文章访问数: 1759
- HTML浏览量: 2
- PDF量: 11
- 被引次数: 0