Paper VIIReading edition

Exact effective dynamics

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

I.9 Exact effective dynamics

I.9.1 Schur–Feshbach reduction

Let \(P\) be a retained packet projector and \(Q=I-P\). In packet coordinates,

\[ H= \begin{pmatrix} H_{PP}&H_{PQ}\\ H_{QP}&H_{QQ} \end{pmatrix}. \]

Theorem I.9.1: Exact Feshbach resolvent

Assume that both \(zI_Q-H_{QQ}\) and \(zI-H\) are invertible. Define

\[ \boxed{ H_{\mathrm{eff}}(z) = H_{PP} +H_{PQ}(z-H_{QQ})^{-1}H_{QP}. } \]

Then

\[ \boxed{ P(z-H)^{-1}P = \left(z-H_{\mathrm{eff}}(z)\right)^{-1} } \]

where the left side is restricted to \(P{\mathcal H}\), the right side is an inverse on that retained space, and both are written in the same orthonormal basis. Here \(z-H_{QQ}\) and \(z-H_{\mathrm{eff}}\) abbreviate the appropriately sized scalar matrices minus the indicated blocks.

Proof of Theorem I.9.1

The block in the lower right corner of \(z-H\) is \(z-H_{QQ}\). Gaussian block elimination produces the Schur complement

\[ z-H_{PP}-H_{PQ}(z-H_{QQ})^{-1}H_{QP}. \]

The assumed invertibility of this block and of the full matrix implies that the Schur complement is invertible. The block in the upper left corner of the full inverse is the inverse of that Schur complement, which is exactly the displayed operator.

The second premise cannot be dropped: at \(z=0\) and \(H=\operatorname{diag}(0,-1)\), the eliminated block is \(1\) but \(zI-H\) is singular. This counterexample is also included in the Lean module.

The self-energy

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

contains the complete influence of the eliminated packet. Bare truncation drops \(\Sigma_P(z)\) and is therefore exact only when the coupling between packets vanishes or the self-energy cancels for another stated reason.

Exact Feshbach reduction has far smaller resolvent error than bare truncation.
Exact Feshbach reduction versus bare truncation.
Text description

A logarithmic bar chart compares projected resolvent error. The exact Feshbach bar is near numerical precision, while the bare-truncation bar is substantially larger. The comparison retains the influence of the eliminated coordinates in one finite example.

Open original figure (PDF)

View full-size figure (new tab)

Norms of four packet-resolved Green-function blocks.
Norms of packet Green-function blocks.
Text description

A two-by-two heatmap places source parity on columns and target parity on rows. The two diagonal blocks have positive displayed norms; the off-diagonal blocks are zero in this example. The color scale represents block Frobenius norms.

Open original figure (PDF)

View full-size figure (new tab)

Numerical certificate I.9.2: Effective resolvent

In the reported test with a random Hermitian matrix of dimension six, the exact Feshbach residual was \(3.2291\times10^{-17}\). The corresponding residual for bare truncation was \(1.3091\times10^{-2}\).

I.9.2 The hidden packet as memory

For a time-independent finite Hamiltonian \(H\) and fixed projectors \(P,Q\), write \(\psi=x+y\), with \(x=P\psi\) and \(y=Q\psi\). Then

\[ \begin{aligned} i\dot x&=H_{PP}x+H_{PQ}y,\\ i\dot y&=H_{QP}x+H_{QQ}y. \end{aligned} \]

Solving the second equation gives

\[ y(t) =e^{-itH_{QQ}}y(0) -i\int_0^t e^{-i(t-s)H_{QQ}}H_{QP}x(s)\,ds. \]

Substitution yields the exact Volterra equation

\[ \boxed{ \begin{aligned} i\dot x(t) ={}&H_{PP}x(t) +H_{PQ}e^{-itH_{QQ}}y(0)\\ &-i\int_0^t K(t-s)x(s)\,ds, \end{aligned} } \]

with memory kernel

\[ \boxed{ K(t)=H_{PQ}e^{-itH_{QQ}}H_{QP}. } \]

The influence of the eliminated degrees of freedom is therefore retained in a self-energy that depends on frequency, or equivalently in an exact memory term. For forward evolution the integral is causal. When the Hamiltonian depends on time, the corresponding propagator between two times replaces the stationary exponential used here. This Volterra derivation is not a separate entry in the finite Lean register.

Source and conversion record

Source SHA-256: f4fb9d650bb8f6a7a9b6ccdd04882b4c165566f81c962f4765983b0ce77714b9

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 VII

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