Dilation situation.

This commit is contained in:
Bokuan Li
2026-09-15 09:55:31 -04:00
parent 1a3c88aa49
commit f020a2037d
7 changed files with 234 additions and 2 deletions
+8
View File
@@ -293,3 +293,11 @@
eprint = {https://math.stackexchange.com/q/150258}, eprint = {https://math.stackexchange.com/q/150258},
url = {https://math.stackexchange.com/q/150258} url = {https://math.stackexchange.com/q/150258}
} }
@book{Paulsen,
title={Completely bounded maps and operator algebras},
author={Paulsen, Vern I and others},
volume={78},
year={2002},
publisher={Cambridge University Press}
}
+51
View File
@@ -34,14 +34,65 @@
\begin{definition}[Transitive] \begin{definition}[Transitive]
\label{definition:tds-transitive} \label{definition:tds-transitive}
Let $G \curvearrowright X$ be a $G$-flow and $x \in X$, then $x$ is \textbf{topologically transitive} with respect to $G$ if $Gx$ is dense in $X$. Let $G \curvearrowright X$ be a $G$-flow and $x \in X$, then $x$ is \textbf{topologically transitive} with respect to $G$ if $Gx$ is dense in $X$.
The flow $G \curvearrowright X$ is \textbf{point transitive} if it admits a topologically transitive point.
Finally, $G \curvearrowright X$ is \textbf{topologically transitive} if for any $U, V \subset X$ non-empty and open, there exists $g \in G$ such that $gU \cap V \ne \emptyset$\footnote{Approximations of points may be translated into approximations of other points. }.
\end{definition} \end{definition}
\begin{proposition}
\label{proposition:tds-point-transitive}
Let $G \curvearrowright X$ be a point transitive $G$-flow, then $G \curvearrowright X$ is also topologically transitive.
\end{proposition}
\begin{proof}
Let $x \in G$ be a topologically transitive point and $U, V \subset X$ be non-empty and open. Since $Gx$ is dense in $G$, $Gx \cap U \ne \emptyset$ and $Gx \cap V \ne \emptyset$. After translating $x$, assume without loss of generality that $x \in U$, then $GU \supset Gx \cap V \ne \emptyset$. Therefore there exists $g \in G$ such that $gU \cap V \ne \emptyset$.
\end{proof}
\begin{definition}[Periodic] \begin{definition}[Periodic]
\label{definition:tds-periodic} \label{definition:tds-periodic}
Let $G \curvearrowright X$ be a $G$-flow and $x \in X$, then $x$ is \textbf{periodic} with respect to $G$ if $Gx$ is finite. Let $G \curvearrowright X$ be a $G$-flow and $x \in X$, then $x$ is \textbf{periodic} with respect to $G$ if $Gx$ is finite.
\end{definition} \end{definition}
\begin{definition}[Minimality]
\label{definition:tds-minimal}
Let $G \curvearrowright X$ be a $G$-flow, then the following are equivalent:
\begin{enumerate}
\item $G \curvearrowright X$ admits no proper subflows.
\item Every point in $X$ is topologically transitive.
\end{enumerate}
If the above holds, then $G \curvearrowright X$ is \textbf{minimal}.
\end{definition}
\begin{proof}
(1) $\Rightarrow$ (2): Let $x \in X$, then $\ol{Gx}$ with the restricted action is a $G$-subflow of $X$. As $\emptyset \ne \ol{Gx}$ and $G \curvearrowright X$ admits no proper subflows, $\ol{Gx} = x$.
(2) $\Rightarrow$ (1): Let $G \curvearrowright Y$ be a $G$-subflow of $G \curvearrowright X$. For any $x \in Y$, $Y \supset \ol{Gx} = X$, so $Y = X$.
\end{proof}
\begin{definition}[Bernoulli Shift]
\label{definition:bernoulli-shift}
Let $G$ be a group, $X = 2^G$ equipped with the product topology, and
\[
\alpha: X \to X \quad \alpha_g(x)(h) = x(g^{-1}h)
\]
Viewing $2^G$ as the power set of $G$, $\alpha_g(A) = gA$ for any $A \subset G$. The $G$-flow $G \curvearrowright X$ is the \textbf{Bernoulli shift} of $G$, and $G$-subflows of $X$ are \textbf{subshifts} of $G$.
\end{definition}
\begin{example}[The Shift]
\label{example:bernoulli-shift-z}
Let $\integer \curvearrowright 2^\integer$ be the Bernoulli shift of $G$, then:
\begin{enumerate}
\item For any $x, y \in 2^\integer$, $y \in \ol{\integer x}$ if and only if for each $A \subset \integer$, there exists $n \in \integer$ such that $y|_A = nx|_A$. In other words, $x$ contains a copy of any segment of $y$\footnote{Topological conditions in Bernoulli shifts translate into combinatorial information. }.
\item $\integer \curvearrowright 2^\integer$ is point transitive, and hence topologically transitive\footnote{It is easier to show topological transitivity: move the trees until they are disjoint}.
\item
\end{enumerate}
\end{example}
+3
View File
@@ -2,3 +2,6 @@
\label{part:logic} \label{part:logic}
This section is plagiarised from This section is plagiarised from
\input{./tp/index.tex}
\input{./notation.tex}
+11
View File
@@ -0,0 +1,11 @@
\chapter{Notations}
\label{chap:logic-notations}
\begin{tabular}{lll}
\textbf{Notation} & \textbf{Description} & \textbf{Source} \\
\hline
% ---- Topological Vector Spaces ----
$\mathcal{L} = (\mathcal{F}, \mathcal{R}, \mathcal{C})$ & Language $\mathcal{L}$ with function symbols $\mathcal{F}$, relation symbols $\mathcal{R}$, and constant symbols $\mathcal{C}$. & \autoref{definition:language} \\
$\mathcal{M} = (M, f^\cm, R^\cm, c^\cm)$ & $\mathcal{L}$-structure with functions $\bracsn{f^\cm|f \in \mathcal{F}}$, relations $\bracsn{R^\cm|R \in \mathcal{R}}$, and constants $\bracsn{c^\cm|c \in \mathcal{C}}$. & \autoref{definition:language-structure}
\end{tabular}
+41
View File
@@ -11,5 +11,46 @@
\end{enumerate} \end{enumerate}
\end{definition} \end{definition}
\begin{definition}[$\mathcal{L}$-Structure]
\label{definition:language-structure}
Let $L = (\mathcal{F}, \mathcal{R}, \mathcal{C})$ be a language, then a \textbf{\mathcal{L}-structure} $\mathcal{M}$ is a family
\[
\mathcal{M} = (M, f^\cm, R^\cm, c^\cm| f \in \mathcal{F}, R \in \mathcal{R}, c \in \mathcal{C})
\]
where
\begin{enumerate}
\item $M$ is a non-empty set, known as the \textbf{universe}, \textbf{domain}, or \textbf{underlying set} of $\mathcal{M}$.
\item For each $f \in \mathcal{F}$, $f^\cm: M^{n_f} \to M$ is a function in $n$-variables in $M$.
\item For each $R \in \mathcal{R}$, $R^\cm \subset M^{n_R}$ is a $n$-ary relation.
\item For each $c \in \mathcal{C}$, $c^\cm \in M$ is a constant.
\end{enumerate}
The assignments $f^\cm$, $R^\cm$, and $c^\cm$ are the \textbf{interpretations} of $(\mathcal{F}, \mathcal{R}, \mathcal{C})$.
\end{definition}
\begin{definition}[Terms]
\label{definition:logic-terms}
Let $\mathcal{L} = (\mathcal{F}, \mathcal{R}, \mathcal{C})$ be a language, then the set of \textbf{$\mathcal{L}$-terms} is the smallest set of expressions $\mathcal{T}(\mathcal{L})$ such that:
\begin{enumerate}
\item $\mathcal{C} \subset \mathcal{T}(\mathcal{L})$.
\item For each $n \in \natp$, the variable symbol $v_n \in \mathcal{T}(\mathcal{L})$.
\item For each $f \in \mathcal{F}$ and $\bracsn{t_j}_1^{n_f} \subset \mathcal{T}(\mathcal{L})$, the expression $f(t_1, \cdots, t_{n_f}) \in \mathcal{T}(\mathcal{L})$.
\end{enumerate}
\end{definition}
\begin{lemma}
\label{lemma:logic-terms-induction}
Let $\mathcal{L} = (\mathcal{F}, \mathcal{R}, \mathcal{C})$ be a language and $\mathcal{T}_0(\mathcal{L}) = \mathcal{C} \cup \seq{v_n}$. For each $n \in \natp$, inductively define
\[
\mathcal{T}_{n}(\mathcal{L}) = \mathcal{T}_{n-1}(\mathcal{L}) \cup \bracsn{f(t_1, \cdots, t_{n_f})|f \in \mathcal{F}, \bracsn{t_j}_1^{n_f} \in \mathcal{T}_{n-1}(\mathcal{L})}
\]
then $\mathcal{T} = \bigcup_{n \in \natz}\mathcal{T}_n$. For each $n \in \natz$, $\mathcal{T}_n$ is the set of \textbf{$\mathcal{L}$-terms of complexity at most $n$}.
\end{lemma}
+120
View File
@@ -81,3 +81,123 @@
\end{proof} \end{proof}
\begin{definition}[Unitary Dilation]
\label{definition:unitary-dilation}
Let $H$ be a complex Hilbert space, $V \in B(H)$ be isometric, $P = I_H - VV^*$ be the projection onto $V(H)^\perp$, and
\[
U = \begin{bmatrix}
V & P \\
0 & V^*
\end{bmatrix}
\]
then:
\begin{enumerate}
\item $U^*U = UU^* = I_{H^2}$. In other words, $U$ is an isometry on $H^2$.
\item Under the identification that $H = H \times \bracsn{0} \subset H^2$, let $Q \in B(H^2)$ be the projection onto $H$, then $V^n = QU^n|_H$ for all $n \ge 0$.
\end{enumerate}
The operator $U \in B(H^2)$ is a \textbf{unitary dilation} of $V$.
\end{definition}
\begin{proof}
(1): By direct computation,
\begin{align*}
UU^* &= \begin{bmatrix}
V & P \\
0 & V^*
\end{bmatrix} \cdot
\begin{bmatrix}
V^* & 0 \\
P & V
\end{bmatrix} = \begin{bmatrix}
VV^* + P & PV \\
V^*P & V^*V
\end{bmatrix} = I_{H^2} \\
U^*U &= \begin{bmatrix}
V^* & 0 \\
P & V
\end{bmatrix} \cdot
\begin{bmatrix}
V & P \\
0 & V^*
\end{bmatrix}
= \begin{bmatrix}
V^*V & V^*P \\
PV & VV^* + P
\end{bmatrix} = I_{H^2} \\
\end{align*}
(2): For each $n \in \natp$,
\[
U^n = \begin{bmatrix}
V & P \\
0 & V^*
\end{bmatrix}^n = \begin{bmatrix}
V^n & \cdots \\
0 & (V^*)^n
\end{bmatrix}
\]
so $QU^n|_{H} = V^n$.
\end{proof}
\begin{definition}[Isometric Dilation]
\label{definition:isometric-dilation}
Let $H$ be a complex Hilbert space, $T \in B(H)$ with $\norm{T}_{B(H)} \le 1$, $D_T = (1 - T^*T)^{1/2}$, and
\[
V: l^2(\natp; H) \to l^2(\natp; H) \quad (Vx)_n = \begin{cases}
Tx_1 &n = 1 \\
D_Tx_1 &n = 2 \\
x_{n-1} &n \ge 3
\end{cases}
\]
then:
\begin{enumerate}
\item $V$ is isometric on $l^2(\natp; H)$.
\item Under the identification that $H = H \oplus 0 \oplus \cdots$, let $Q$ be the orthogonal projection onto $H$, then $T^n = QV^n|_H$ for all $n \ge 0$.
\end{enumerate}
The operator $V$ is an \textbf{isometric dilation} of $T$.
\end{definition}
\begin{proof}
(1): Since $\norm{T}_H \le 1$, $1 - T^*T \ge 0$, and $(1 - T^*T)^{1/2}$ is well-defined. For each $x \in H$,
\begin{align*}
\norm{Tx}_H^2 + \norm{D_Tx}_H^2 &= \dpn{T^*Tx, x}{H} + \dpn{(1 - T^*T)x, x}{H} \\
&= \dpn{x, x}{H} = \norm{x}_H^2
\end{align*}
Thus for each $x = \seq{x_n} \in l^2(H)$,
\begin{align*}
\norm{Vx}_{l^2(\natp; H)}^2 &= \norm{Tx_1}_H^2 + \norm{D_Tx_1}_H^2 + \sum_{n = 3}^\infty \norm{x_{n-1}}_H^2 \\
&= \norm{x}_H^2 + \sum_{n = 2}^\infty \norm{x_{n}}_H^2 = \norm{x}_{l^2(\natp; H)}^2
\end{align*}
\end{proof}
\begin{theorem}[Sz.-Nagy's Dilation Theorem]
\label{theorem:sznagy-dilation}
Let $H$ be a complex Hilbert space and $T \in B(H)$ with $\norm{T}_{B(H)} \le 1$, then there exists a Hilbert space $K \supset H$ and a unitary operator $U \in B(K)$ such that $T^n = P_HU^n|_H$ for all $n \in \natp$, where $P_H$ is the orthogonal projection onto $H$.
\end{theorem}
\begin{proof}[Proof, {{\cite[Theorem 1.1]{Paulsen}}}. ]
Let $K = l^2(\natp; H)^2$ and identify $H = (H \oplus \bracsn{0} \oplus \cdot) \times \bracsn{0}$. Let $V \in B(l^2(\natp; H))$ be the isometric dilation of $T$, and $U \in B(l^2(\natp; H)^2)$ be the unitary dilation of $V$. Identify $l^2(\natp; H)$ as a subspace of $K$ as $l^2(\natp; H) \times \bracsn{0}$, and let $P_{l^2(\natp; H)}$ be the orthogonal projection onto $l^2(\natp; H)$, then for each $n \ge 0$,
\begin{align*}
V^n &= P_{l^2(\natp; H)}U^n|_{l^2(\natp; H)} \\
T^n = P_{H}V^n|_{H} &= P_HU^n|_{H}
\end{align*}
\end{proof}
\begin{corollary}[Von Neumann's Inequality]
\label{corollary:von-neumann-inequality-polynomial}
Let $H$ be a complex Hilbert space and $T \in B(H)$ with $\norm{T}_{B(H)} \le 1$, then for any $p \in \complex[x]$,
\[
\norm{p(T)}_{B(H)} \le \sup_{z \in B_\complex(0, 1)}|p(z)|
\]
\end{corollary}
\begin{proof}[Proof, {{\cite[Corollary 1.2]{Paulsen}}}. ]
By \hyperref[Sz.-Nagy's Dilation Theorem]{theorem:sznagy-dilation}, there exists a Hilbert space $K \supset H$ and a unitary operator $U \in B(K)$ such that $T^n = P_HU^n|_H$ for all $n \in \natp$, where $P_H$ is the orthogonal projection onto $H$.
Thus for any $p \in \complex[x]$, $p(T) = P_Hp(U)|_H$. As $U$ is unitary, the \hyperref[continuous functional calculus]{definition:continuous-functional-calculus} shows that $\norm{p(U)}_{B(K)} \le \sup_{z \in B_\complex(0, 1)}|p(z)|$. Therefore $\norm{p(T)}_{B(H)} \le \sup_{z \in B_\complex(0, 1)}|p(z)|$.
\end{proof}
-2
View File
@@ -82,5 +82,3 @@
\] \]
\end{example} \end{example}