Toeplitz Quantization and Convexity

Let \lambda^m = ( \lambda_0,\cdots,\lambda_m) \in \mathbb{R}^{m+1} be the eigenvalues of a hermitian matrix A . Let diag(A)= ( a_{00}, \cdots, a_{mm}) be the diagonal of A then \lambda^m majorizes diag(A).

Let x \in \mathbb{R}^{m+1}

• ( Rado’s theorem)

  \{ y \in \mathbb{R}^{m+1}, y \prec x \} = co ( \sum\nolimits_m x)

• y \prec x if and only if \sum_{i=0}^{i = m} f(y_i) \leq \sum_{i=0}^{i = m} f(x_i) for any convex function whose domain contains all the numbers x_i, y_i, 0 \leq i \leq m.

Let \lambda^m = ( \lambda_0,\cdots,\lambda_m) \in \mathbb{R}^{m+1}.
Let \mathcal{H}_{\lambda^m } be the set of all hermitian matrices whose spectrum is \lambda^m.
Let p^m :\mathcal{H}_{\lambda^m } \longrightarrow \mathbb{R}^{m+1} be the map that picks the diagonal of a matrix.
Then the image of the map p^m is the convex set co( \sum\nolimits_m{\lambda^m}).

Let L^2(\mathbb{CP}(1)) be the set of square integrable functions on the sphere. Let P: L^2(\mathbb{CP}(1)) \longrightarrow L^2[0,1] be the projection \displaystyle{ P(f) = \frac{1}{2 \pi}\int_0^{2 \pi} f(z,\theta)d\theta}.

Then f^* majorizes P( f).

Let f \in L^2([0,1]). The set \displaystyle{\Omega(f) = \{g \in L^2([0,1]), g \prec f \}} is weakly compact and convex. Its set of extreme points is
\displaystyle{ \{ f^* \circ \phi\; |\; \phi \; \mbox{ is a measure preserving transformation of }\; [0,1] \}}.

Let \sum_{m} be the symmetric group of (m+1) letters. There exists a one to one group homomorphism \Psi^m :\sum_{m} \longrightarrow \emph{Imeas(I)} that sends \sigma to \widehat{\sigma} and if we identify \sum_m with \Psi^m(\sum_m), then \bigcup_m\sum_m is a dense set in Smeas(I) for the strong operator topology.

Let \lambda^m = \left(\lambda_0^m , \lambda_1^m , \cdots ,\lambda_m^m \right) be the eigenvalues of T_f^m arranged in non-increasing order and let \Lambda^m (s) be the real step function defined on the interval [0,1[ by

Then the sequence \left(\Lambda^m(s)\right)_m converges point-wise almost everywhere to the decreasing rearrangement f^*(s) of the function f.

Given a smooth real valued function f on \mathbb{CP}(1), let T^m_f be Toeplitz quantization of f and let \mu^m be its spectral measure . Then \dfrac{\mu^m}{m+1} tends weakly to a limiting measure as m tends to infinity, this limiting measure being

Let (X,\Sigma,\mu) be a finite measure space,and f,\,f_n:\,X \rightarrow \mathbb{R} be a sequence of measurable functions such that f_n converge in distribution to f. Then on the Lebesgue measure space ( I, \mathcal{B},\left\lvert \cdot \right\rvert), where I = (0, \mu(X), \mathcal{B} is the Borel \sigma-algebra of I, and \left \lvert \cdot \right\rvert is the Lebesgue measure, there exists measurable functions g_n,g: I \rightarrow \mathbb{R} such that g_n(t) \rightarrow g(t) \;a.e [\left\lvert \cdot \right\rvert], and

x \in \mathbb{R}, n \geq 1.

Let f \in C^{\infty}(\mathbb{CP}(1)). Let f^* be the decreasing rearrangement of f. Let T_f^m be the Toeplitz quantization of f and let \lambda^m = (\lambda^m_0,\cdots,\lambda^m_m) be the eigenvalues of T_f^m. Let \Omega(f^*) = \{ g \in L^2([0,1], g \prec f^*\}. Then the closed convex hull of the set Smeas(I) \cdot f^* = \{f^* \circ \phi \;| \phi \; \in Smeas(I)\} is the weak closure of the topological lim sup of the convex sets E_m = p^m( co(\sum_m \cdot \lambda^m)).