| − | 那么如何用有限的、来自现实的数据,去逼近[[无限]]维的Koopman线性[[算子]]呢?最开始的方法是'''[[动力学模式分解]](DMD,Dynamic Mode Decomposition)'''<ref name=Schmid_2008>P. J. Schmid and J. Sesterhenn, Dynamic mode decomposition of numerical and experimental data, in 61st Annual Meeting of the APS Division of Fluid Dynamics, American Physical Society, 2008. (Cited on pp. 249, 258)</ref><ref name=Schmid_2010>P. J. Schmid, Dynamic mode decomposition of numerical and experimental data, J. Fluid Mech., 656 (2010), pp. 5--28. (Cited on pp. 231, 237, 238, 249, 250, 255, 258, 284, 292)</ref>,它直接把“[[状态变量]]本身”当作观测函数。然而,对于真正的[[非线性系统]],DMD的近似就太粗糙了。故此后学者们围绕着“如何更有效地拓展观测空间”与“如何优化求解过程”展开。'''扩展动态[[模式分解]](EDMD, Extended Dynamic Mode Decomposition)'''<ref name=william_2015>M. O. Williams, I. G. Kevrekidis, and C. W. Rowley, A data-driven approximation of the Koopman operator: Extending dynamic mode decomposition, J. Nonlinear Sci., 6 (2015), pp. 1307--1346. (Cited on pp. 246, 247, 256, 284, 285, 292)</ref>通过预定义的[[非线性]]函数显式拓展;'''残差动态[[模式分解]](ResDMD, Residual DMD)'''<ref name=Colbrook_2023>Colbrook, M. J., Ayton, L. J. & Szőke, M. Residual dynamic mode decomposition: Robust and verified koopmanism. J. Fluid Mech. 955, A21 (2023).</ref><ref name=Colbrook_2024>Colbrook, M. J. & Townsend, A. Rigorous data‐driven computation of spectral properties of koopman operators for dynamical systems. Commun. Pure Appl. Math. 77, 221–283 (2024).</ref>致力于严格评估逼近精度并筛选可靠模态;'''核动态[[模式分解]](Kernel DMD)'''<ref name=william_2015_kernal>M. O. Williams, C. W. Rowley, and I. G. Kevrekidis, A kernel approach to data-driven Koopman spectral analysis, J. Comput. Dynam., 2 (2015), pp. 247--265. (Cited on pp. 255, 256, 284, 285, 294, 318)</ref><ref name=Gonzalez_2023_kernal>Gonzalez, E., Abudia, M., Jury, M., Kamalapurkar, R. & Rosenfeld, J. A. The kernel perspective on dynamic mode decomposition. Preprint at https://doi.org/10.48550/arXiv.2106.00106 (2023).</ref>利用[[核技巧]]隐式构建观测函数空间;'''基于[[神经网络]]的动态[[模式分解]]([[Neural Network]] DMD)'''<ref name=CW2018>C. Wehmeyer and F. Noe\', Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics, J. Chem. Phys., 148 (2018), pp. 1--9. (Cited on pp. 284, 290, 291)</ref><ref name=NT2017>N. Takeishi, Y. Kawahara, and T. Yairi, Learning Koopman invariant subspaces for dynamic mode decomposition, in Advances in Neural Information Processing Systems, 2017, pp. 1130--1140. (Cited on pp. 284, 290)</ref>用[[深度学习]]自动学习最优的拓展方式;'''时间延迟动态[[模式分解]](Time-delay DMD/Hankel DMD)'''<ref name=brunton_2017>S. L. Brunton, B. W. Brunton, J. L. Proctor, E. Kaiser, and J. N. Kutz, Chaos as an intermittently forced linear system, Nature Commun., 8 (2017), pp. 1--9. (Cited on pp. 257, 275, 284, 287, 288, 289, 294)</ref>通过历史数据构造[[嵌入空间]];'''[[随机]]动态[[模式分解]](Randomized DMD)'''<ref name=ranDMD>Erichson, N. B., Mathelin, L., Kutz, J. N. & Brunton, S. L. Randomized dynamic mode decomposition. Siam J. Appl. Dyn. Syst. 18, 1867–1891 (2019)</ref>专注于利用[[随机]]算法提升高维问题的计算效率;'''最优[[模式分解]](OPD/Optimal DMD)'''<ref name=opdmd>T. Askham and J. N. Kutz, Variable projection methods for an optimized dynamic mode decomposition, SIAM J. Appl. Dyn. Syst., 17 (2018), pp. 380--416, https://doi.org/10. 1137/M1124176. (Cited on pp. 253, 254)</ref>将问题重构为全局[[非线性]]优化以求取最佳拟合;'''多分辨率动态[[模式分解]](mrDMD, Multiresolution DMD)'''<ref name=mrdmd>J. N. Kutz, X. Fu, and S. L. Brunton, Multiresolution dynamic mode decomposition, SIAM J. Appl. Dyn. Syst., 15 (2016), pp. 713--735, https://doi.org/10.1137/15M1023543. (Cited on pp. 257, 278)</ref>则能分离系统在不同时间尺度或层次上的[[动力学]];'''前向-后向动态[[模式分解]](fbDMD, forward-backward DMD)'''<ref name=fbdmd>S. T. Dawson, M. S. Hemati, M. O. Williams, and C. W. Rowley, Characterizing and correcting for the effect of sensor noise in the dynamic mode decomposition, Exp. Fluids, 57 (2016), pp. 1--19. (Cited on p. 253)</ref>通过平均正反向数据提升稳定性;'''张量动态[[模式分解]](Tensor DMD)'''<ref name=tensordmd>S. Klus, P. Gel{\ss}, S. Peitz, and C. Sch\"utte, Tensor-based dynamic mode decomposition, Nonlinearity, 31 (2018), pp. 3359--3380. (Cited on p. 257)</ref>处理具有空间结构的多维数据;'''受控动态[[模式分解]](Controlled DMD)'''<ref name=DMDc>J. L. Proctor, S. L. Brunton, and J. N. Kutz, Dynamic mode decomposition with control, SIAM J. Appl. Dyn. Syst., 15 (2016), pp</ref>则旨在辨识含外部输入系统的Koopman模型。DMD算法现在已经扩展出许多不同的版本,在这里只是罗列的部分情况,下文会对主流的几种方法进行更详细的介绍,其他算法请参考文献。 | + | 那么如何用有限的、来自现实的数据,去逼近[[无限]]维的Koopman线性[[算子]]呢?最开始的方法是'''[[动力学模式分解]](DMD,Dynamic Mode Decomposition)'''<ref name=Schmid_2008>P. J. Schmid and J. Sesterhenn, Dynamic mode decomposition of numerical and experimental data, in 61st Annual Meeting of the APS Division of Fluid Dynamics, American Physical Society, 2008.</ref><ref name=Schmid_2010>P. J. Schmid, Dynamic mode decomposition of numerical and experimental data, J. Fluid Mech., 656 (2010), pp. 5--28.</ref>,它直接把“[[状态变量]]本身”当作观测函数。然而,对于真正的[[非线性系统]],DMD的近似就太粗糙了。故此后学者们围绕着“如何更有效地拓展观测空间”与“如何优化求解过程”展开。'''扩展动态[[模式分解]](EDMD, Extended Dynamic Mode Decomposition)'''<ref name=william_2015>M. O. Williams, I. G. Kevrekidis, and C. W. Rowley, A data-driven approximation of the Koopman operator: Extending dynamic mode decomposition, J. Nonlinear Sci., 6 (2015), pp. 1307--1346. (Cited on pp. 246, 247, 256, 284, 285, 292)</ref>通过预定义的[[非线性]]函数显式拓展;'''残差动态[[模式分解]](ResDMD, Residual DMD)'''<ref name=Colbrook_2023>Colbrook, M. J., Ayton, L. J. & Szőke, M. Residual dynamic mode decomposition: Robust and verified koopmanism. J. Fluid Mech. 955, A21 (2023).</ref><ref name=Colbrook_2024>Colbrook, M. J. & Townsend, A. Rigorous data‐driven computation of spectral properties of koopman operators for dynamical systems. Commun. Pure Appl. Math. 77, 221–283 (2024).</ref>致力于严格评估逼近精度并筛选可靠模态;'''核动态[[模式分解]](Kernel DMD)'''<ref name=william_2015_kernal>M. O. Williams, C. W. Rowley, and I. G. Kevrekidis, A kernel approach to data-driven Koopman spectral analysis, J. Comput. Dynam., 2 (2015), pp. 247--265. (Cited on pp. 255, 256, 284, 285, 294, 318)</ref><ref name=Gonzalez_2023_kernal>Gonzalez, E., Abudia, M., Jury, M., Kamalapurkar, R. & Rosenfeld, J. A. The kernel perspective on dynamic mode decomposition. Preprint at https://doi.org/10.48550/arXiv.2106.00106 (2023).</ref>利用[[核技巧]]隐式构建观测函数空间;'''基于[[神经网络]]的动态[[模式分解]]([[Neural Network]] DMD)'''<ref name=CW2018>C. Wehmeyer and F. Noe\', Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics, J. Chem. Phys., 148 (2018), pp. 1--9. (Cited on pp. 284, 290, 291)</ref><ref name=NT2017>N. Takeishi, Y. Kawahara, and T. Yairi, Learning Koopman invariant subspaces for dynamic mode decomposition, in Advances in Neural Information Processing Systems, 2017, pp. 1130--1140. (Cited on pp. 284, 290)</ref>用[[深度学习]]自动学习最优的拓展方式;'''时间延迟动态[[模式分解]](Time-delay DMD/Hankel DMD)'''<ref name=brunton_2017>S. L. Brunton, B. W. Brunton, J. L. Proctor, E. Kaiser, and J. N. Kutz, Chaos as an intermittently forced linear system, Nature Commun., 8 (2017), pp. 1--9. (Cited on pp. 257, 275, 284, 287, 288, 289, 294)</ref>通过历史数据构造[[嵌入空间]];'''[[随机]]动态[[模式分解]](Randomized DMD)'''<ref name=ranDMD>Erichson, N. B., Mathelin, L., Kutz, J. N. & Brunton, S. L. Randomized dynamic mode decomposition. Siam J. Appl. Dyn. Syst. 18, 1867–1891 (2019)</ref>专注于利用[[随机]]算法提升高维问题的计算效率;'''最优[[模式分解]](OPD/Optimal DMD)'''<ref name=opdmd>T. Askham and J. N. Kutz, Variable projection methods for an optimized dynamic mode decomposition, SIAM J. Appl. Dyn. Syst., 17 (2018), pp. 380--416, https://doi.org/10. 1137/M1124176. (Cited on pp. 253, 254)</ref>将问题重构为全局[[非线性]]优化以求取最佳拟合;'''多分辨率动态[[模式分解]](mrDMD, Multiresolution DMD)'''<ref name=mrdmd>J. N. Kutz, X. Fu, and S. L. Brunton, Multiresolution dynamic mode decomposition, SIAM J. Appl. Dyn. Syst., 15 (2016), pp. 713--735, https://doi.org/10.1137/15M1023543. (Cited on pp. 257, 278)</ref>则能分离系统在不同时间尺度或层次上的[[动力学]];'''前向-后向动态[[模式分解]](fbDMD, forward-backward DMD)'''<ref name=fbdmd>S. T. Dawson, M. S. Hemati, M. O. Williams, and C. W. Rowley, Characterizing and correcting for the effect of sensor noise in the dynamic mode decomposition, Exp. Fluids, 57 (2016), pp. 1--19. (Cited on p. 253)</ref>通过平均正反向数据提升稳定性;'''张量动态[[模式分解]](Tensor DMD)'''<ref name=tensordmd>S. Klus, P. Gel{\ss}, S. Peitz, and C. Sch\"utte, Tensor-based dynamic mode decomposition, Nonlinearity, 31 (2018), pp. 3359--3380.</ref>处理具有空间结构的多维数据;'''受控动态[[模式分解]](Controlled DMD)'''<ref name=DMDc>J. L. Proctor, S. L. Brunton, and J. N. Kutz, Dynamic mode decomposition with control, SIAM J. Appl. Dyn. Syst., 15 (2016), pp</ref>则旨在辨识含外部输入系统的Koopman模型。DMD算法现在已经扩展出许多不同的版本,在这里只是罗列的部分情况,下文会对主流的几种方法进行更详细的介绍,其他算法请参考文献。 |
| − | 除了利用[[神经网络]]来学习Koopman嵌入之外,[[Koopman理论]]也被应用于理解[[神经网络]]<ref name=IM2020>I. Manojlovic\', M. Fonoberova, R. Mohr, A. Andrejc\v uk, Z. Drmac\v , Y. Kevrekidis, and I. Mezic\', Applications of Koopman Mode Analysis to Neural Networks, preprint, https://arxiv.org/abs/2006.11765, 2020. (Cited on p. 290)</ref><ref>A. S. Dogra and W. T. Redman, Optimizing neural networks via Koopman operator theory, in Advances in Neural Information Processing Systems, H. Larochelle, M. Ranzato, R. Hadsell, M. F. Balcan, and H. Lin, eds., Curran Associates, 2020, pp. 2087--2097. (Cited on p. 290)</ref>以及更广泛的算法<ref>F. Dietrich, T. N. Thiem, and I. G. Kevrekidis, On the Koopman operator of algorithms, SIAM J. Appl. Dyn. Syst., 19 (2020), pp. 860--885, https://doi.org/10.1137/19M1277059. (Cited on p. 290)</ref>。 | + | 除了利用[[神经网络]]来学习Koopman嵌入之外,[[Koopman理论]]也被应用于理解[[神经网络]]<ref name=IM2020>I. Manojlovic\', M. Fonoberova, R. Mohr, A. Andrejc\v uk, Z. Drmac\v , Y. Kevrekidis, and I. Mezic\', Applications of Koopman Mode Analysis to Neural Networks, preprint, https://arxiv.org/abs/2006.11765, 2020.</ref><ref>A. S. Dogra and W. T. Redman, Optimizing neural networks via Koopman operator theory, in Advances in Neural Information Processing Systems, H. Larochelle, M. Ranzato, R. Hadsell, M. F. Balcan, and H. Lin, eds., Curran Associates, 2020, pp. 2087--2097.</ref>以及更广泛的算法<ref>F. Dietrich, T. N. Thiem, and I. G. Kevrekidis, On the Koopman operator of algorithms, SIAM J. Appl. Dyn. Syst., 19 (2020), pp. 860--885, https://doi.org/10.1137/19M1277059.</ref>。 |
| − | 另外在实践中,对于具有混沌性质的动力系统还需要注意动力学封闭性的问题,仅靠升维可能是不够的。Champion, Brunton 和 Kutz等人表明<ref name=champion_2019>K. P. Champion, S. L. Brunton, and J. N. Kutz, Discovery of nonlinear multiscale systems: Sampling strategies and embeddings, SIAM J. Appl. Dyn. Syst., 18 (2019), pp. 312--333, https://doi.org/10.1137/18M1188227. (Cited on pp. 288, 289)</ref>,对于[[混沌系统]],即使有近似Koopman不变的测量系统,在识别封闭的线性模型方面仍存在挑战。无论多么高维的线性模型,都无法捕捉多个[[不动点]]或具有正[[Lyapunov指数]]的[[混沌]]所特有的不可预测行为<ref name=Brunton_2016>S. L. Brunton, B. W. Brunton, J. L. Proctor, and J. N. Kutz, Koopman invariant subspaces and finite linear representations of nonlinear dynamical systems for control, PLoS ONE, 11 (2016), art. e0150171. (Cited on pp. 234, 247, 248, 254, 284, 285, 286, 288, 295, 308, 309, 312, 318)</ref>。与其构建一个封闭的线性模型,不如在前 <math>p-1</math> 个变量上构建一个线性模型,并将最后一个变量强加为一个外力项<ref name=brunton_2017 />。 | + | 另外在实践中,对于具有混沌性质的动力系统还需要注意动力学封闭性的问题,仅靠升维可能是不够的。Champion, Brunton 和 Kutz等人表明<ref name=champion_2019>K. P. Champion, S. L. Brunton, and J. N. Kutz, Discovery of nonlinear multiscale systems: Sampling strategies and embeddings, SIAM J. Appl. Dyn. Syst., 18 (2019), pp. 312--333, https://doi.org/10.1137/18M1188227.</ref>,对于[[混沌系统]],即使有近似Koopman不变的测量系统,在识别封闭的线性模型方面仍存在挑战<ref name=mezic_2022>Mezić, I. (2022). On Numerical Approximations of the Koopman Operator. Mathematics, 10(7), 1180. https://doi.org/10.3390/math10071180</ref>。无论多么高维的线性模型,都无法捕捉多个[[不动点]]或具有正[[Lyapunov指数]]的[[混沌]]所特有的不可预测行为<ref name=Brunton_2016>S. L. Brunton, B. W. Brunton, J. L. Proctor, and J. N. Kutz, Koopman invariant subspaces and finite linear representations of nonlinear dynamical systems for control, PLoS ONE, 11 (2016), art. e0150171.</ref><ref name-mezic_2022 />。与其构建一个封闭的线性模型,不如在前 <math>p-1</math> 个变量上构建一个线性模型,并将最后一个变量强加为一个外力项<ref name=brunton_2017 />。 |
| | 现代Koopman[[算子]]理论为在[[非线性系统]]中解释这些经典的基于Hankel方法的结果提供了一个有价值的新视角。在 Hankel 矩阵上计算 DMD 的做法最早由Tu等人<ref name=Tu_2014 />提出,并由Brunton等人<ref name=brunton_neural>B. W. Brunton, L. A. Johnson, J. G. Ojemann, and J. N. Kutz, Extracting spatial-- temporal coherent patterns in large-scale neural recordings using dynamic mode decomposition, J. Neurosci. Methods, 258 (2016), pp. 1--15, https://doi.org/10.1016/j.jneumeth. 2015.10.010. (Cited on pp. 237, 238, 257, 260, 261, 278, 289)</ref>用于[[神经科学]]领域。Hankel矩阵与Koopman[[算子]]之间的联系,由 Brunton等人<ref name=brunton_2017 />在[[HAVOK]]框架中建立。随后的一些工作为该方法提供了更完善的理论基础<ref name=arbabi_2017>H. Arbabi and I. Mezic\', Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the Koopman operator, SIAM J. Appl. Dyn. Syst., 16 (2017), pp. 2096--2126, https://doi.org/10.1137/17M1125236. (Cited on pp. 257, 275, 289, 292, 294)</ref><ref name=das_2019>S. Das and D. Giannakis, Delay-coordinate maps and the spectra of Koopman operators, J. Statist. Phys., 175 (2019), pp. 1107--1145. (Cited on pp. 257, 289, 293)</ref>。[[HAVOK]] 方法也常被称为延迟DMD<ref name=Tu_2014 />或Hankel-DMD<ref name=arbabi_2017 />。 | | 现代Koopman[[算子]]理论为在[[非线性系统]]中解释这些经典的基于Hankel方法的结果提供了一个有价值的新视角。在 Hankel 矩阵上计算 DMD 的做法最早由Tu等人<ref name=Tu_2014 />提出,并由Brunton等人<ref name=brunton_neural>B. W. Brunton, L. A. Johnson, J. G. Ojemann, and J. N. Kutz, Extracting spatial-- temporal coherent patterns in large-scale neural recordings using dynamic mode decomposition, J. Neurosci. Methods, 258 (2016), pp. 1--15, https://doi.org/10.1016/j.jneumeth. 2015.10.010. (Cited on pp. 237, 238, 257, 260, 261, 278, 289)</ref>用于[[神经科学]]领域。Hankel矩阵与Koopman[[算子]]之间的联系,由 Brunton等人<ref name=brunton_2017 />在[[HAVOK]]框架中建立。随后的一些工作为该方法提供了更完善的理论基础<ref name=arbabi_2017>H. Arbabi and I. Mezic\', Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the Koopman operator, SIAM J. Appl. Dyn. Syst., 16 (2017), pp. 2096--2126, https://doi.org/10.1137/17M1125236. (Cited on pp. 257, 275, 289, 292, 294)</ref><ref name=das_2019>S. Das and D. Giannakis, Delay-coordinate maps and the spectra of Koopman operators, J. Statist. Phys., 175 (2019), pp. 1107--1145. (Cited on pp. 257, 289, 293)</ref>。[[HAVOK]] 方法也常被称为延迟DMD<ref name=Tu_2014 />或Hankel-DMD<ref name=arbabi_2017 />。 |