m |
|||
(47 intermediate revisions by 5 users not shown) | |||
Line 1: | Line 1: | ||
[[Category:method]] |
[[Category:method]] |
||
+ | [[Category:Differential_algebraic_system]] |
||
− | The moment-matching methods are also called the Krylov subspace methods, as well as |
||
+ | [[Category:linear]] |
||
− | Pad<math>\'e</math> approximation methods. These methods are applicable for non-parametric linear time invariant systems, often the descriptor systems, e.g. |
||
+ | [[Category:first differential order]] |
||
− | <math>Insert formula here</math> |
||
+ | [[Category:linear algebra]] |
||
+ | ==Description== |
||
+ | The moment-matching methods are also called the ''Krylov'' subspace methods<ref name="freund03"/>, as well as |
||
+ | ''Padé'' approximation methods<ref name="feldmann95"/>. They belong to the [[Projection based MOR]] methods. These methods are applicable to non-parametric linear time invariant systems, often descriptor systems, e.g. |
||
+ | :<math>E \dot{x}(t)=A x(t)+B u(t),</math> |
||
− | They are very efficient in many engineering applications, including circuit simulation, where the robustness of the methods were originally found. |
||
+ | :<math>y(t)=Cx(t).</math> |
||
+ | They are very efficient in many engineering applications, such as circuit simulation, Microelectromechanical systems (MEMS) simulation, etc.. |
||
− | The basic steps are as follows. First, the transfer function~(\ref{eq:TFM}) is expanded into a power series at an expansion point $s_0\in\mathbb{C}\cup \infty$. |
||
+ | |||
− | % |
||
+ | The basic steps are as follows. First, the transfer function |
||
− | % |
||
+ | |||
− | Let $s=s_0+\sigma$, then, within the convergence radius of the series, we have |
||
+ | :<math>H(s)=Y(s)/U(s)=C(sE-A)^{-1}B</math> |
||
− | \[ |
||
+ | |||
− | \begin{array}{lll} |
||
+ | is expanded into a power series at an expansion point <math>s_0\in\mathbb{C}\cup \infty</math>. |
||
− | &&H(s_0 + \sigma)= L^T[(s_{0}+\sigma){I}-A]^{-1}B \\ |
||
+ | |||
− | &&=L^T[\sigma { I}+(s_{0}{ I}-{ A})]^{-1}B\\ |
||
+ | Let <math>s=s_0+\sigma</math>, then, within the convergence radius of the series, we have |
||
− | &&=L^T[{ I}-\sigma(s_0{ I}-{ A})^{-1}]^{-1}[-(s_0{ I}-{ A})]^{-1}B\\ |
||
+ | |||
− | &&=L^T[{ I}+\sigma(s_0{ I}- A )^{-1}+\sigma^2[(s_0{ I}-{ A})^{-1}]^{2}+\ldots]\times\\ |
||
+ | :<math>H(s_0 + \sigma)= C[(s_{0}+\sigma){E}-A]^{-1}B</math> |
||
− | &&\quad({ A}-s_0{I})^{-1}B\\ |
||
+ | |||
− | &&=\sum \limits^\infty_{i=0}\underbrace{L^T[(s_0{ I}-{A})^{-1}]^i({ A}-s_0{ I})^{-1}B}_{:= m_i(s_0)} \, \sigma^i, |
||
+ | ::<math>=C[\sigma { E}+(s_{0}{ E}-{ A})]^{-1}B</math> |
||
− | \end{array} |
||
+ | |||
− | \] |
||
+ | ::<math>=C[{ I}+\sigma(s_0{ E}-{ A})^{-1}E]^{-1}(s_0{ E}-{ A})^{-1}B</math> |
||
− | % |
||
+ | |||
− | % |
||
+ | ::<math>=C[{ I}-\sigma(s_0{ E}- A )^{-1}E+\sigma^2[(s_0{ E}-{ A})^{-1}E]^{2}+\ldots] |
||
− | where $m_i(s_0)$ are called the moments of the transfer function about $s_0$ for $i=0,1,2,\ldots$. |
||
+ | (s_0{E}-{ A})^{-1}B</math> |
||
− | If the expansion point is chosen as zero then the moments simplify to $m_i(0)=L^\mathrm{T}(-A^{-1})^{i+1}B$. |
||
+ | |||
− | For $s_0=\infty$ the moments are also called Markov parameters which can be computed by $L^\mathrm{T} \bA^{i-1}B$. |
||
+ | ::<math>=\sum \limits^\infty_{i=0}\underbrace{C[-(s_0{ E}-{A})^{-1}E]^i(s_0{ E}-{ A})^{-1}B}_{:= m_i(s_0)} \, \sigma^i,</math> |
||
+ | |||
+ | where <math>m_i(s_0)</math> are called the moments of the transfer function about <math>s_0</math> for <math>i=0,1,2,\ldots</math>. |
||
+ | If the expansion point is chosen as zero, then the moments simplify to <math>m_i(0)=C(A^{-1}E)^i(-A^{-1}B)</math>. |
||
The goal in moment-matching model reduction is the construction of a reduced order |
The goal in moment-matching model reduction is the construction of a reduced order |
||
− | system where some moments |
+ | system where some moments <math>\hat m_i</math> of the associated transfer function <math>\hat H</math> match some moments |
− | of the original transfer function |
+ | of the original transfer function <math>H</math>. |
+ | |||
− | A few important classes of approximations are listed in Table~\ref{tab:moments}. |
||
+ | The matrices <math>V</math> and <math>W</math> for model order reduction can be computed |
||
− | % |
||
+ | from the vectors which are associated with the moments, for |
||
− | % |
||
+ | example, using a single expansion point <math>s_0=0</math>, by |
||
− | \begin{center} |
||
+ | |||
− | \begin{table*}[ht] |
||
+ | :<math>\textrm{range}(V)=\textrm{span}\{\tilde B,({ A}^{-1}E)^2 \tilde B, \ldots,({ A}^{-1}E)^{r}{\tilde B}\}, \quad \quad \quad \quad \quad \quad \quad \quad\quad \quad \quad \ (1) </math> |
||
− | \hfill{} |
||
+ | |||
− | \begin{tabular}{l|ll} |
||
+ | :<math>\textrm{range}(W)=\textrm{span}\{C^T, E^T{ A}^{-T}C^T,(E^T{A}^{-T})^2C^T, \ldots |
||
− | %\hline |
||
+ | ,(E^T{A}^{-T})^{r-1}C^T\}, \quad \quad (2) </math> |
||
− | Name of reduced order system& Matched moments &\\\hline |
||
+ | |||
− | %\cline{3-4} |
||
+ | where <math>\tilde B=-A^{-1}B</math>. |
||
− | %\hline |
||
+ | |||
− | Pad\'e approximation~\cite{Bak75} & $m_i(s_0) = \hat m_i(s_0)$, & $i=0,1,\cdots,2r-1$ \\ |
||
+ | The transfer function <math>\hat H</math> of the reduced model has good approximation properties around <math>s_0</math>, which matches the first <math>2r</math> moments of <math>H(s)</math> at <math>s_0</math>. |
||
− | Partial realization~\cite{GraL83} & $m_i(\infty) = \hat m_i(\infty)$, & $i=0,1,\cdots,2r-1$ \\ |
||
+ | |||
− | Multipoint Pad\'e approximation or & $m_i(s_j) = \hat m_i(s_j)$, & $i=0,1,\cdots,2r_j-1$, for $j=1,\cdots,k$, and $r_1+\dots+r_k = r$ \\ |
||
+ | Using a set of <math>k</math> distinct expansion points <math>\{s_1,\cdots,s_k\}</math>, the reduced model obtained by, e.g., |
||
− | rational interpolation~\cite{AndA90,Bak75} & & |
||
+ | |||
− | %\hline |
||
+ | |||
− | \end{tabular} |
||
+ | :<math>\textrm{range}(V)=\textrm{span}\{(A-s_1 {E})^{-1}E\tilde B,\ldots,(A-s_k {E})^{-1}E\tilde B \}, \quad \quad \quad \quad \quad \quad \quad \quad (3)</math> |
||
− | \hfill{} |
||
+ | |||
− | \caption{Some examples for model reduction by moment-matching.} |
||
+ | :<math>\textrm{range}(W)=\textrm{span}\{E^T(A-s_1 {E})^{-T}C^T,\ldots,E^T(A-s_k {E})^{-T}C^T \},\quad \quad \quad \quad \quad (4) </math> |
||
− | \label{tab:moments} |
||
+ | |||
− | \end{table*} |
||
+ | matches the first two moments at each <math>s_j</math>, <math>j=1,\ldots,k</math>, see <ref name="grimme97"/>. The reduced model is in the form as below |
||
− | \end{center} |
||
+ | |||
+ | :<math>W^TE \frac{d{V z}}{t}=W^TA Vz +W^T B u(t), \quad \hat{y}(t)=CVz.</math> |
||
+ | |||
+ | For the case of one expansion point in (1)(2), it can be seen that the columns of <math>V</math>, <math>W</math> span Krylov subspaces |
||
+ | which can easily be computed by Arnoldi or Lanczos methods. The matrices <math>V</math> and <math>W</math> in (3)(4) can be computed with the rational Krylov algorithm in<ref name="grimme97"/> or with the modified Gram-Schmidt process. In these algorithms only a few number of linear systems need to be solved, where matrix-vector multiplications are only used if using iterative solvers, which are simple to implement and the complexity of the resulting |
||
+ | methods is roughly <math>O(n r^2)</math> for sparse matrices <math>A, E</math>. |
||
+ | |||
+ | ==References== |
||
+ | |||
+ | <references> |
||
+ | |||
+ | <ref name="freund03">R.W. Freund, "<span class="plainlinks">[http://dx.doi.org/10.1017/S0962492902000120 Model reduction methods based on Krylov subspaces]</span>". Acta Numerica, 12:267-319, 2003.</ref> |
||
+ | |||
+ | <ref name="feldmann95">P. Feldmann and R.W. Freund, "<span class="plainlinks">[http://dx.doi.org/10.1109/43.384428 Efficient linear circuit analysis by Pade approximation via the Lanczos process]</span>". IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst., 14:639-649, 1995.</ref> |
||
+ | |||
+ | <ref name="grimme97">E.J. Grimme, "<span class="plainlinks">[http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.19.9254&rep=rep1&type=pdf Krylov projection methods for model reduction]</span>. PhD thesis, Univ. Illinois, Urbana-Champaign, 1997.</ref> |
||
+ | |||
+ | </references> |
Latest revision as of 14:55, 25 August 2022
Description
The moment-matching methods are also called the Krylov subspace methods[1], as well as Padé approximation methods[2]. They belong to the Projection based MOR methods. These methods are applicable to non-parametric linear time invariant systems, often descriptor systems, e.g.
They are very efficient in many engineering applications, such as circuit simulation, Microelectromechanical systems (MEMS) simulation, etc..
The basic steps are as follows. First, the transfer function
is expanded into a power series at an expansion point .
Let , then, within the convergence radius of the series, we have
where are called the moments of the transfer function about
for
.
If the expansion point is chosen as zero, then the moments simplify to
.
The goal in moment-matching model reduction is the construction of a reduced order
system where some moments of the associated transfer function
match some moments
of the original transfer function
.
The matrices and
for model order reduction can be computed
from the vectors which are associated with the moments, for
example, using a single expansion point
, by
where .
The transfer function of the reduced model has good approximation properties around
, which matches the first
moments of
at
.
Using a set of distinct expansion points
, the reduced model obtained by, e.g.,
matches the first two moments at each ,
, see [3]. The reduced model is in the form as below
For the case of one expansion point in (1)(2), it can be seen that the columns of ,
span Krylov subspaces
which can easily be computed by Arnoldi or Lanczos methods. The matrices
and
in (3)(4) can be computed with the rational Krylov algorithm in[3] or with the modified Gram-Schmidt process. In these algorithms only a few number of linear systems need to be solved, where matrix-vector multiplications are only used if using iterative solvers, which are simple to implement and the complexity of the resulting
methods is roughly
for sparse matrices
.
References
- ↑ R.W. Freund, "Model reduction methods based on Krylov subspaces". Acta Numerica, 12:267-319, 2003.
- ↑ P. Feldmann and R.W. Freund, "Efficient linear circuit analysis by Pade approximation via the Lanczos process". IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst., 14:639-649, 1995.
- ↑ 3.0 3.1 E.J. Grimme, "Krylov projection methods for model reduction. PhD thesis, Univ. Illinois, Urbana-Champaign, 1997.