Paper VIReading edition

Schur Feshbach elimination

Reading tools
Reading libraryFind statementsInspect Lean
Switch work

Saved reading

Read side by side

Start reading

Resources & source

Original TeX with line links · Download TeX

The wording and mathematics follow the retained manuscript. This reading format does not add a proof review or extend Lean coverage.

Skip contents and start reading

E14 Schur Feshbach elimination

Let \(X=X_P\oplus X_Q\) and write

\[ A= \begin{pmatrix} A_{PP}&A_{PQ}\\ A_{QP}&A_{QQ} \end{pmatrix}. \]

For \(zI-A_{QQ}\) invertible, define

\[ \Sigma_P(z) = A_{PQ}(zI-A_{QQ})^{-1}A_{QP} \]

and

\[ F_P(z) = zI-A_{PP}-\Sigma_P(z). \]

Theorem E14.1: Schur Feshbach resolvent

Assume \(zI-A_{QQ}\) is invertible. Then \(zI-A\) is invertible if and only if \(F_P(z)\) is invertible. In that case,

\[ P(zI-A)^{-1}P = F_P(z)^{-1}. \]

In finite dimension,

\[ \det(zI-A) = \det(zI-A_{QQ})\det F_P(z). \]

Proof of Theorem E14.1

Set

\[ L_z= \begin{pmatrix} I&-A_{PQ}(zI-A_{QQ})^{-1}\\ 0&I \end{pmatrix} \]

and

\[ S_z= \begin{pmatrix} F_P(z)&0\\ -A_{QP}&zI-A_{QQ} \end{pmatrix}. \]

Direct multiplication gives \(zI-A=L_zS_z\). The first factor is invertible. The claims follow from triangular inversion and the determinant of the second factor. This is the finite block form of the Feshbach reduction [Feshbach1958].

E14.1 Time domain memory

For the homogeneous equation \(\dot x=Ax\), variation of constants gives

\[ x_Q(t) = {\mathrm e}^{tA_{QQ}}x_Q(0) + \int_0^t {\mathrm e}^{(t-s)A_{QQ}}A_{QP}x_P(s) {\,\mathrm d} s. \]

Substitution into the retained equation gives

\[ \begin{aligned} \dot x_P(t) ={}& A_{PP}x_P(t) + A_{PQ}{\mathrm e}^{tA_{QQ}}x_Q(0)\\ &+ \int_0^t {\mathcal K}(t-s)x_P(s) {\,\mathrm d} s, \end{aligned} \]

where

\[ {\mathcal K}(t) = A_{PQ}{\mathrm e}^{tA_{QQ}}A_{QP}. \]

Theorem E14.2: Memory and self energy

If \(z\) lies to the right of the spectrum of \(A_{QQ}\), then

\[ \int_0^\infty {\mathrm e}^{-zt}{\mathcal K}(t){\,\mathrm d} t = \Sigma_P(z). \]

If, for constants \(M\geq0\) and \(\gamma>0\),

\[ {\left\lVert{\mathrm e}^{tA_{QQ}}\right\rVert} \leq M{\mathrm e}^{-\gamma t}, \]

then

\[ {\left\lVert{\mathcal K}(t)\right\rVert} \leq M{\left\lVert A_{PQ}\right\rVert}{\left\lVert A_{QP}\right\rVert}{\mathrm e}^{-\gamma t} \]

and

\[ {\left\lVert{\mathcal K}\right\rVert}_{L^1(0,\infty)} \leq \frac{M}{\gamma} {\left\lVert A_{PQ}\right\rVert}{\left\lVert A_{QP}\right\rVert}. \]

Proof of Theorem E14.2

If \(X_Q=\{0\}\), the memory kernel and self-energy are both zero, so the identity is immediate. Otherwise write \(B=A_{QQ}\), which is a bounded operator on the hidden Banach space, and put \(s(B)=\max\operatorname{Re}{\operatorname{Spec}}(B)\). Choose a real \(\beta\) with \(s(B)<\beta<\operatorname{Re}z\). Spectral mapping and the spectral-radius formula give \(r({\mathrm e}^B)={\mathrm e}^{s(B)}<{\mathrm e}^\beta\), hence \({\left\lVert{\mathrm e}^{nB}\right\rVert}\leq M_\beta{\mathrm e}^{\beta n}\) for every integer \(n\geq0\), after increasing \(M_\beta\) to cover the finitely many initial values. Write \(t=n+r\), where \(0\leq r<1\). Then

\[ {\left\lVert{\mathrm e}^{tB}\right\rVert}\leq M_\beta\exp({\left\lVert B\right\rVert}+\max(0,-\beta)){\mathrm e}^{\beta t}. \]

Thus \({\mathrm e}^{-zt}{\mathrm e}^{tB}\) is Bochner integrable and tends to zero. Integrating its derivative gives both inverse identities for \(R=\int_0^\infty{\mathrm e}^{-zt}{\mathrm e}^{tB}{\,\mathrm d} t\), so \(R=(zI-B)^{-1}\). This is the bounded-generator growth principle in [EngelNagel2000], Chapter IV, Corollary 3.12. Multiplication by the coupling blocks gives the first identity. The remaining estimates follow directly from the assumed positive decay rate.

Theorem E14.3: Criterion for exact local closure

Assume \(x_Q(0)=0\) and \(d=\dim X_Q<\infty\). The following conditions are equivalent.

  1. For every retained initial state,

    \[ P{\mathrm e}^{tA}P = {\mathrm e}^{tA_{PP}}P \]

    for all \(t\geq0\).

  2. \(\Sigma_P(z)=0\) for every sufficiently large \(z\).

  3. \({\mathcal K}(t)=0\) for every \(t\geq0\).

  4. If \(d=\dim X_Q\), then

    \[ A_{PQ}A_{QQ}^kA_{QP}=0 \]

    for \(0\leq k<d\).

Proof of Theorem E14.3

By Theorem E2.5, the Laplace transform of the projected propagator is the projected resolvent. By Theorem E14.1, the first condition is equivalent to

\[ (zI-A_{PP}-\Sigma_P(z))^{-1} = (zI-A_{PP})^{-1} \]

for all sufficiently large \(z\), which is equivalent to the second condition. The second and third conditions are equivalent by Theorem E14.2 and uniqueness of the Laplace transform.

For large \(z\),

\[ \Sigma_P(z) = \sum_{k=0}^{\infty} \frac{A_{PQ}A_{QQ}^kA_{QP}}{z^{k+1}}. \]

Thus the second condition implies the moment identities. Conversely, the first \(d\) identities imply all later ones by the Cayley Hamilton theorem, so the series vanishes.

Remark E14.4

A finite hidden block produces a strictly proper rational self energy. It tends to zero as \(z\) tends to infinity. Hence a nonzero constant self energy cannot arise from hidden dynamics alone. A constant correction must already be included in the retained generator.

Source and conversion record

Source SHA-256: 736e615131f57daf341b6cb0600761d58950450947c85d433206148cca228333

Unsupported prose commands stop the build. The mathematical expression sequence is checked against the source, with manuscript macros expanded for display.

Conversion records

Notation in Paper VI

Look up a symbol without leaving this chapter. Select text before opening help to search for it.

Search this work’s notation.

    Symbols can have different meanings in different passages. Read each result in its own context.

    Open the full notation index