
<p>Front </p><p></p><ul style="display: flex;"><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>Front page </p><p>The transfer operator approach to “quantum chaos” </p><p>Classical mechanics and the Laplace-Beltrami operator on PSL<sub style="top: 0.12em;">2</sub>(Z)\H </p><p>Tobias Mu¨hlenbruch <br>Joint work with <a href="/goto?url=http://www.dynamik.tu-clausthal.de/institute/" target="_blank">D. Mayer </a>and <a href="/goto?url=http://www.mathematik.tu-darmstadt.de/fbereiche/AlgGeoFA/staff/stroemberg.php" target="_blank">F. Str</a><a href="/goto?url=http://www.mathematik.tu-darmstadt.de/fbereiche/AlgGeoFA/staff/stroemberg.php" target="_blank">¨</a><a href="/goto?url=http://www.mathematik.tu-darmstadt.de/fbereiche/AlgGeoFA/staff/stroemberg.php" target="_blank">omberg </a></p><p><a href="/goto?url=http://www.math.tu-clausthal.de/" target="_blank">Institut f</a><a href="/goto?url=http://www.math.tu-clausthal.de/" target="_blank">u</a><a href="/goto?url=http://www.math.tu-clausthal.de/" target="_blank">¨r Mathematik </a><br><a href="/goto?url=http://www.tu-clausthal-de" target="_blank">TU Clausthal </a></p><p><a href="mailto:[email protected]" target="_blank">[email protected] </a></p><p>22 January 2009, <a href="/goto?url=http://www.fernuni-hagen.de/" target="_blank">Fernuniversit</a><a href="/goto?url=http://www.fernuni-hagen.de/" target="_blank">¨</a><a href="/goto?url=http://www.fernuni-hagen.de/" target="_blank">at in Hagen </a></p><p>Front </p><p></p><ul style="display: flex;"><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>Outline of the presentation </p><p>1</p><p>Modular surface </p><p>23456</p><p>Spectral theory Geodesics and coding Transfer operator Connection to spectral theory Conclusions </p><p>Front </p><p>Modular surface </p><p></p><ul style="display: flex;"><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>The full modular group PSL (Z) <br>2</p><p>Let PSL<sub style="top: 0.12em;">2</sub>(Z) be the full modular group </p><p>PSL<sub style="top: 0.12em;">2</sub>(Z) = hS, T|(ST)<sup style="top: -0.35em;">3 </sup>= 1i = SL<sub style="top: 0.12em;">2</sub>(Z) mod { 1} </p><p></p><ul style="display: flex;"><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li></ul><p></p><p>01</p><ul style="display: flex;"><li style="flex:1">−1 </li><li style="flex:1">1</li></ul><p>0<br>11<br>10<br>01</p><ul style="display: flex;"><li style="flex:1">with S = </li><li style="flex:1">, T = </li><li style="flex:1">and 1 = </li><li style="flex:1">.</li></ul><p>0</p><p></p><ul style="display: flex;"><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li><li style="flex:1">ꢂ</li></ul><p></p><p>az+b cz+d </p><p></p><ul style="display: flex;"><li style="flex:1">a</li><li style="flex:1">b</li></ul><p></p><p>if z ∈ C, </p><p></p><ul style="display: flex;"><li style="flex:1">M¨obius transformations: </li><li style="flex:1">z = </li></ul><p></p><p>a</p><p></p><ul style="display: flex;"><li style="flex:1">c</li><li style="flex:1">d</li></ul><p></p><p>if z = ∞. </p><p>c</p><p>Three orbits: the upper halfplane H = {x + iy; y > 0}, the projective real line P<sup style="top: -0.3em;">1</sup><sub style="top: 0.23em;">R </sub>and the lower half plane. z<sub style="top: 0.12em;">1</sub>, z<sub style="top: 0.12em;">2 </sub>are PSL<sub style="top: 0.12em;">2</sub>(Z)-equivalent if ∃M ∈ PSL<sub style="top: 0.12em;">2</sub>(Z) with </p><p>M z<sub style="top: 0.12em;">1 </sub>= z<sub style="top: 0.12em;">2</sub>. </p><p>The full modular group PSL<sub style="top: 0.12em;">2</sub>(Z) is generated by<sub style="top: 0.67em;">−1 </sub></p><ul style="display: flex;"><li style="flex:1">translation T : z → z + 1 and inversion S : z → </li><li style="flex:1">.</li></ul><p></p><p>z</p><p>ꢃ</p><p>12</p><p>(Closed) fundamental domain F = z ∈ H; |z| ≥ 1, |Re(z)| ≤ </p><p>.</p><p>Front </p><p>Modular surface </p><p></p><ul style="display: flex;"><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>The modular surface PSL (Z)\H <br>2</p><p>The upper half plane H = {x + iy; y > 0} can be viewed as a hyperbolic plane with constant negative curvature −1. </p><p>dx<sup style="top: -0.26em;">2</sup>+dy<sup style="top: -0.26em;">2 </sup></p><p>Line element ds </p><p>ds<sup style="top: -0.31em;">2 </sup>= </p><p>y<sup style="top: -0.16em;">2 </sup></p><p>dxdy </p><p>Volume element dA </p><p>dA = </p><p>y<sup style="top: -0.16em;">2 </sup></p><p>The M¨obius transformations are compatible with the hyperbolic metric. This way, it makes sense to speak of the </p><p>modular surface PSL<sub style="top: 0.12em;">2</sub>(Z)\H. </p><p>Gluing F along the identified edges, we obtain a realization of the modular surface PSL<sub style="top: 0.12em;">2</sub>(Z)\H, a non-compact, finite surface with <sub style="top: 0.35em;">√</sub>one </p><p>1+i <br>2<br>3</p><p>cusp at z → i∞ and two conic singularities at z = i and z = The hyperbolic plane H is also called the Poincar´e half plane. PSL<sub style="top: 0.12em;">2</sub>(Z)\H is arithmetic ( connection to number theory). <br>.</p><p>Front </p><p>Modular surface </p><p></p><ul style="display: flex;"><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>Hecke triangle groups </p><p>Possible extension of PSL<sub style="top: 0.12em;">2</sub>(Z): </p><p>Hecke triangle groups G<sub style="top: 0.12em;">q </sub></p><p>G<sub style="top: 0.12em;">q </sub>= hS, T<sub style="top: 0.12em;">q</sub>|(ST<sub style="top: 0.12em;">q</sub>)<sup style="top: -0.3em;">q </sup>= 1i with <br>, λ<sub style="top: 0.12em;">q </sub>= 2 cos π/q and q = 3, 4, 5, . . .. </p><p></p><ul style="display: flex;"><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li></ul><p></p><ul style="display: flex;"><li style="flex:1">ꢄ</li><li style="flex:1">ꢅ</li></ul><p></p><p>10</p><p>λ<sub style="top: 0.12em;">q </sub></p><p>1</p><p>T = </p><p>Example: PSL<sub style="top: 0.12em;">2</sub>(Z) = G<sub style="top: 0.12em;">3</sub>. M¨obius transformation extends to G<sub style="top: 0.12em;">q</sub>. Fundamental domain </p><p></p><ul style="display: flex;"><li style="flex:1">n</li><li style="flex:1">o</li></ul><p></p><p>λ<sub style="top: 0.08em;">q </sub></p><p>F<sub style="top: 0.12em;">q </sub>= z ∈ H; |z| ≥ 1, |Re(z)| ≤ </p><p>.</p><p>2</p><p>Fundamental domain forms a (2, q, ∞) </p><p>π</p><p>2</p><p>π</p><p>q</p><p></p><ul style="display: flex;"><li style="flex:1">hyperbolic triangle with angles </li><li style="flex:1">,</li><li style="flex:1">and 0. </li></ul><p>G<sub style="top: 0.12em;">q</sub>\H is finite and non-compact. G<sub style="top: 0.12em;">q</sub>\H is non-arithmetic for q = 3, 4, 6. </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li></ul><p></p><p>Spectral theory </p><p></p><ul style="display: flex;"><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>Maass cusp forms </p><p>Maass cusp form u </p><p>u : H → C real-analytic function, </p><p></p><ul style="display: flex;"><li style="flex:1">ꢄ</li><li style="flex:1">ꢅ</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">∆u = s(1 − s) u with ∆ = −y<sup style="top: -0.3em;">2 </sup>∂<sub style="top: 0.21em;">x</sub><sup style="top: -0.3em;">2 </sup>+ ∂<sub style="top: 0.21em;">y</sub><sup style="top: -0.3em;">2 </sup></li><li style="flex:1">,</li></ul><p>u(M z) = u(z) for all M ∈ PSL<sub style="top: 0.12em;">2</sub>(Z), </p><p></p><ul style="display: flex;"><li style="flex:1">ꢄ</li><li style="flex:1">ꢅ</li></ul><p></p><p>u(x + iy) = O y<sup style="top: -0.3em;">C </sup>as y → ∞ for all C ∈ R. </p><p>Maass cusp form at </p><p>1</p><p>s = + i13.7797 . . .. </p><p>∆ admits self-adjoined extension in L<sup style="top: -0.3em;">2 </sup>PSL<sub style="top: 0.12em;">2</sub>(Z)\H . </p><p>2</p><p></p><ul style="display: flex;"><li style="flex:1">ꢄ</li><li style="flex:1">ꢅ</li></ul><p></p><p>s is called spectral parameter. </p><p>s(1 − s) > <sup style="top: -0.33em;">1</sup><sub style="top: 0.28em;">4 </sub>, i.e., s ∈ + iR<sup style="top: -0.31em;">⋆</sup>. </p><p>12</p><p>Discrete spectrum, eigenvalues have finite multiplicity, It is assumed that eigenvalues have multiplicity 1. Precise location of eigenvalues is unknown. </p><p>P</p><p>√</p><p>(2π |n| y) e<sup style="top: -0.3em;">2πinx </sup></p><p>Maass cusp form at </p><p>u(x + iy) = </p><p></p><ul style="display: flex;"><li style="flex:1">y</li><li style="flex:1">a<sub style="top: 0.12em;">n </sub>K </li></ul><p></p><p>1</p><p>s− <sub style="top: 0.2em;">2 </sub></p><p>0=n∈Z </p><p>12</p><p>s = + i9.533 . . .. </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li></ul><p></p><p>Spectral theory </p><p></p><ul style="display: flex;"><li style="flex:1">Geodesics and coding </li><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>The Laplace-Beltrami operator </p><p></p><ul style="display: flex;"><li style="flex:1">ꢄ</li><li style="flex:1">ꢅ</li></ul><p></p><p>L<sup style="top: -0.31em;">2 </sup>PSL<sub style="top: 0.12em;">2</sub>(Z)\H are realized by functions h satisfying h: H → C measurable, h(M z) = f (z) for all M ∈ PSL<sub style="top: 0.12em;">2</sub>(Z) and </p><p>R</p><p>2</p><p><sub style="top: 0.3em;">F </sub>|h(x, iy)| y<sup style="top: -0.3em;">−2</sup>dxdy < ∞. </p><p>Some properties of ∆ </p><p>The Laplace-Beltrami operator on L<sup style="top: -0.31em;">2</sup>(M) is the self-adjoined extension of ∆. </p><p>Spectrum of ∆ is discrete. </p><p>Area(H) </p><p>4π </p><p>Eigenvalues λ = s(1 − s) obey Weyl’s law: ♯{λ ≤ Λ} ∼ Conjectures about Eigenvalue statistics. Quantum unique ergodicity. <br>.<br>Arithmetic properties: Hecke operators, associated L-series satisfy a GRH. </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li></ul><p></p><p>Geodesics and coding </p><p></p><ul style="display: flex;"><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>Hurwitz continued fractions </p><p>Definition (Hurwitz continued fractions (CF)) </p><p>We identify a sequence of integers, a<sub style="top: 0.08em;">0 </sub>∈ Z, and a<sub style="top: 0.08em;">1</sub>, a<sub style="top: 0.08em;">2</sub>, . . . ∈ Z<sup style="top: -0.32em;">⋆ </sup>with a point <br>−1 </p><p></p><ul style="display: flex;"><li style="flex:1">a</li><li style="flex:1">a</li></ul><p></p><p>2</p><p>x = T<sup style="top: -0.36em;">a</sup><sup style="top: -0.26em;">0 </sup>ST <sup style="top: -0.26em;">1 </sup>ST · · · 0 = a<sub style="top: 0.08em;">0 </sub>+ </p><p>=: [a0; a1, a2, . . .] </p><p>−1 </p><p>a<sub style="top: 0.08em;">1 </sub>+ </p><p>−1 </p><p>... </p><p>a</p><p>2</p><p>+</p><p>and say that it is a </p><p>non-regular (formal) CF, [a0; a1, a2, . . .] in general. </p><p>regular CF, [a0; a1, a2, . . .], if it does not contain “forbidden blocks”: </p><ul style="display: flex;"><li style="flex:1">no 1 appear </li><li style="flex:1">and </li><li style="flex:1">if a<sub style="top: 0.08em;">i </sub>= 2 then a<sub style="top: 0.08em;">i+1 </sub>≶ 0. </li></ul><p></p><p>π = [3; −7, 16, 294, 3, 4, 5, 15, . . .] and e = [3; 4, 2, −5, −2, 7, 2, −9, . . .] </p><p>Equivalent points </p><p>x and y are equivalent :⇔ there exist a g ∈ PSL<sub style="top: 0.08em;">2</sub>(Z) such that gx = y ⇔ the CF of x and y have the same tail or the CF of x and y have tail [ 3 ] and [ −3 ]. </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li></ul><p></p><p>Geodesics and coding </p><p></p><ul style="display: flex;"><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>Associated dynamical system </p><p>The generating map f </p><p>12</p><p></p><ul style="display: flex;"><li style="flex:1">(x) ∈ Z denotes the nearest integer of x, i.e. |x − (x)| ≤ </li><li style="flex:1">.</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">ꢆ</li><li style="flex:1">ꢇ</li></ul><p></p><p>1</p><p>For I = −<sub style="top: 0.28em;">2 </sub>, <sup style="top: -0.33em;">1</sup><sub style="top: 0.28em;">2 </sub>the generating map for the CF of x is </p><p></p><ul style="display: flex;"><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">−1 </li><li style="flex:1">−1 </li></ul><p></p><p>f : I → I; x → </p><p>−</p><p>.</p><p></p><ul style="display: flex;"><li style="flex:1">x</li><li style="flex:1">x</li></ul><p></p><p>The coefficients a<sub style="top: 0.12em;">0</sub>, a<sub style="top: 0.12em;">1</sub>, . . . computed by </p><p>−1 </p><p>a<sub style="top: 0.12em;">0 </sub>= (x) and x<sub style="top: 0.12em;">n+1 </sub>= f (x<sub style="top: 0.12em;">n</sub>) = </p><p>− a<sub style="top: 0.12em;">n+1 </sub></p><p>x<sub style="top: 0.08em;">n </sub></p><p>satisfy x = [a<sub style="top: 0.12em;">0</sub>; a<sub style="top: 0.12em;">1</sub>, . . .] and the CF is regular. </p><p>Natural extension of f </p><p>The natural extension of f is </p><p></p><ul style="display: flex;"><li style="flex:1">ꢀ</li><li style="flex:1">ꢁ</li></ul><p></p><p>−1 <br>Ω → Ω; (x, y) → f (x), </p><p>y + a<sub style="top: 0.12em;">1 </sub></p><p>with x = [0; a<sub style="top: 0.12em;">1</sub>, . . .]. </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li></ul><p></p><p>Geodesics and coding </p><p></p><ul style="display: flex;"><li style="flex:1">Transfer operator </li><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>PSL (Z) and geodesics <br>2</p><p>Recall PSL<sub style="top: 0.12em;">2</sub>(Z) = hS, Ti. We denote geodesics γ on H by its base points: γ = (γ , γ<sub style="top: 0.12em;">+</sub>). </p><p>−</p><p>In the diagram we illustrate the closed geodesic γ = ([0; −3, −4], [0; −4, −3]<sup style="top: -0.3em;">−1</sup>). </p><p>Theorem </p><p>Each geodesic γ<sup style="top: -0.31em;">′ </sup>is PSL<sub style="top: 0.12em;">2</sub>(Z)-equivalent with a geodesic </p><p>γ = (γ , γ<sub style="top: 0.12em;">+</sub>) satisfying (γ , γ<sup style="top: -0.35em;">−1</sup>) ∈ Ω. </p><p></p><ul style="display: flex;"><li style="flex:1">−</li><li style="flex:1">−</li></ul><p></p><p>+</p><p>If γ<sup style="top: -0.3em;">′ </sup>is closed then γ = [0; a<sub style="top: 0.12em;">1</sub>, . . . , a<sub style="top: 0.12em;">n</sub>] is regular and </p><p>−</p><p>γ<sub style="top: 0.2em;">+</sub><sup style="top: -0.35em;">−1 </sup>= [0; a<sub style="top: 0.12em;">n</sub>, . . . , a<sub style="top: 0.12em;">1</sub>] </p><p>If γ and υ satisfy (γ , γ<sup style="top: -0.36em;">−1</sup>), (υ , υ<sup style="top: -0.36em;">−1</sup>) ∈ Ω and no base point is </p><p></p><ul style="display: flex;"><li style="flex:1">−</li><li style="flex:1">−</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">+</li><li style="flex:1">+</li></ul><p></p><p>√<br>3− </p><p>5</p><p></p><ul style="display: flex;"><li style="flex:1">PSL<sub style="top: 0.12em;">2</sub>(Z)-equivalent with [0; 3] = </li><li style="flex:1">then γ = υ. </li></ul><p></p><p>2</p><p>Nearest λ-multiple continued fractions </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li></ul><p></p><p>Transfer operator </p><p></p><ul style="display: flex;"><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>The Ising model and transfer matrices </p><p>Ernst Ising (1900 – 1998) discussed an 1 − d lattice spin model. </p><p>Ising-Model </p><p>Config. space S = { 1}<sup style="top: -0.32em;">N</sup>; left-shift τ : S → S, (τξ)<sub style="top: 0.08em;">i </sub>:= ξ<sub style="top: 0.08em;">i+1</sub>. </p><p></p><ul style="display: flex;"><li style="flex:1">P</li><li style="flex:1">P</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">∞</li><li style="flex:1">∞</li></ul><p></p><p>Total energy: E = −J <sub style="top: 0.23em;">i=1 </sub>ξ<sub style="top: 0.08em;">i </sub>ξ<sub style="top: 0.08em;">i+1 </sub>+ B <sub style="top: 0.23em;">i=1 </sub>ξ<sub style="top: 0.08em;">i </sub>with spin interaction J and magnetic field interaction B.) </p><p>Ernst Ising ≈ 1925. </p><p>Partition function Z<sub style="top: 0.08em;">m</sub>(A, s) </p><p>P</p><p>X</p><p>m−1 k=0 </p><p>k</p><p>A(τ ξ) </p><p>e<sup style="top: -0.36em;">−s </sup></p><p></p><ul style="display: flex;"><li style="flex:1">with s = </li><li style="flex:1">and A(ξ) = J ξ<sub style="top: 0.08em;">0</sub>ξ<sub style="top: 0.08em;">1 </sub>+ Bξ<sub style="top: 0.08em;">0</sub>. </li></ul><p></p><p>1Temp. </p><p>Z<sub style="top: 0.08em;">m</sub>(A, s) = </p><p>ξ∈S; m−periodic </p><p> free energy = lim<sub style="top: 0.08em;">m→∞ </sub>m<sup style="top: -0.31em;">−1 </sup>log Z<sub style="top: 0.08em;">m</sub>(A, s). <br>Ising rewrote Z<sub style="top: 0.08em;">m </sub>as </p><p>X</p><p>e(ξ<sub style="top: 0.08em;">1</sub>, ξ<sub style="top: 0.08em;">2</sub>) · · · e(ξ<sub style="top: 0.08em;">m</sub>, ξ<sub style="top: 0.08em;">1</sub>) with e(ξ<sub style="top: 0.08em;">i </sub>, ξ<sub style="top: 0.08em;">j </sub>) = e<sup style="top: -0.36em;">−s(Jξ ξ −Bξ ) </sup></p><p>.</p><p></p><ul style="display: flex;"><li style="flex:1">i</li><li style="flex:1">j</li><li style="flex:1">i</li></ul><p></p><p>Z<sub style="top: 0.08em;">m</sub>(A, s) = </p><p>ξ</p><p>1</p><p>,...,ξ ∈{ 1} </p><p>m</p><p>Transfer matrix L<sub style="top: 0.08em;">s </sub></p><p></p><ul style="display: flex;"><li style="flex:1">„</li><li style="flex:1">«</li></ul><p></p><ul style="display: flex;"><li style="flex:1">`</li><li style="flex:1">´</li></ul><p></p><p>e(+1, +1) e(+1, −1) e(−1, +1) e(−1, −1) </p><ul style="display: flex;"><li style="flex:1">L<sub style="top: 0.08em;">s </sub>:= </li><li style="flex:1">satisfies Z<sub style="top: 0.08em;">m</sub>(A, s) = trace L<sub style="top: 0.13em;">s</sub><sup style="top: -0.32em;">m </sup></li><li style="flex:1">.</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li></ul><p></p><p>Transfer operator </p><p></p><ul style="display: flex;"><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>General form of a transfer operator </p><p>General form of a transfer operator </p><p>Given a set Λ and maps f : Λ → Λ and g : Λ → C, a transfer operator L acting on functions h: Λ → C is defined by </p><p>X</p><ul style="display: flex;"><li style="flex:1">ꢄ</li><li style="flex:1">ꢅ</li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">Lh (x) = </li><li style="flex:1">g(y) h(y) </li></ul><p></p><p>y∈f <sup style="top: -0.16em;">−1</sup>(x) </p><p>Remarks: <br>Usually, take g = |J|<sup style="top: -0.42em;">−1 </sup>if the Jacobian J of f exists. </p><p>ꢄ</p><p>ꢅ<sub style="top: 0.17em;">−1 </sub></p><p>P</p><p>L of the form Lh(x) = </p><p>f <sup style="top: -0.3em;">′</sup>(y) </p><p>h(y) is also known as a </p><p>y∈f <sup style="top: -0.17em;">−1</sup>(x) </p><p>Perron-Frobenius Operator. </p><p>more </p><p>Relation of L to the dynamical zeta-function: <br>1ζ(z) = </p><p>.</p><p>det<sup style="top: -0.37em;">c </sup>(1 − zL) </p><p></p><ul style="display: flex;"><li style="flex:1">Front </li><li style="flex:1">Modular surface </li><li style="flex:1">Spectral theory </li><li style="flex:1">Geodesics and coding </li></ul><p></p><p>Transfer operator </p><p></p><ul style="display: flex;"><li style="flex:1">Connection to spectral theory </li><li style="flex:1">Conclusions </li></ul><p></p><p>The transfer operator for continued fractions </p><p>Recall respectively define: </p><p>0 −1 <br>1 0 </p><p></p><ul style="display: flex;"><li style="flex:1">`</li><li style="flex:1">´</li><li style="flex:1">`</li><li style="flex:1">´</li></ul><p></p><p>1 1 0 1 </p><p>S = </p><p></p><ul style="display: flex;"><li style="flex:1">and T = </li><li style="flex:1">∈ PSL<sub style="top: 0.08em;">2</sub>(Z), </li></ul><p></p><p></p><ul style="display: flex;"><li style="flex:1">˛</li><li style="flex:1">`</li><li style="flex:1">`</li><li style="flex:1">´´ </li><li style="flex:1">`</li><li style="flex:1">´</li><li style="flex:1">`</li><li style="flex:1">´</li></ul><p></p><p>−s a b </p><p>c d </p><p>(z) := (cz + d)<sup style="top: -0.32em;">2 </sup></p><p>h</p><p>and </p><p>az+b cz+d </p><p>˛</p><p>h</p><p>2s </p><p></p><ul style="display: flex;"><li style="flex:1">ˆ</li><li style="flex:1">˜</li><li style="flex:1">ˆ</li><li style="flex:1">˜</li></ul><p></p><p>f : −<sup style="top: -0.3em;">1</sup><sub style="top: 0.26em;">2 </sub>, </p><p>→ −<sup style="top: -0.3em;">1</sup><sub style="top: 0.26em;">2 </sub>, </p><p>;</p><p>x → </p><p>(mod 1) = T<sup style="top: -0.32em;">−n</sup>S x for some n ∈ Z. </p><p>12<br>12<br>−1 </p><p>x</p><p>The associated transfer operator is formally given by </p><p>ꢄ</p><p>ꢅ<sub style="top: 0.17em;">−s </sub></p><p>P</p><p>L<sub style="top: 0.12em;">s </sub>h(x) := </p><p>f <sup style="top: -0.3em;">′</sup>(y) </p><p></p><ul style="display: flex;"><li style="flex:1">h(y) </li><li style="flex:1">(x ∈ [−1/2, 1/2]) </li></ul><p></p><p>y∈f <sup style="top: -0.17em;">−1</sup>(x) </p><p>Banach space (with sup-norm) V := C(D) ∩ C<sup style="top: -0.3em;">ω</sup>(D<sup style="top: -0.3em;">◦</sup>), D = {z; |z| ≤ 1}. </p><p>Transfer operator for continued fractions </p><p>The operator L<sub style="top: 0.12em;">s </sub>: : V × V → V × V , Re(s) > 1, is defined as </p><p></p><ul style="display: flex;"><li style="flex:1">ꢈ</li><li style="flex:1">ꢈ</li></ul><p>ꢀP<sub style="top: 0.21em;">∞ </sub></p><p>P<sub style="top: 0.2em;">∞ </sub></p><p>ꢁ</p><p><sub style="top: 0.25em;">n=3 </sub>h<sub style="top: 0.12em;">1</sub><sub style="top: 0.33em;">ꢈ2s</sub>ST<sup style="top: -0.3em;">n </sup></p><p>++</p><p><sub style="top: 0.25em;">n=2 </sub>h<sub style="top: 0.12em;">2</sub><sub style="top: 0.33em;">ꢈ2s </sub>ST<sup style="top: -0.3em;">−n </sup></p><p></p><ul style="display: flex;"><li style="flex:1">ꢈ</li><li style="flex:1">ꢈ</li></ul><p></p><p>~</p><p></p><ul style="display: flex;"><li style="flex:1">P<sub style="top: 0.2em;">∞ </sub></li><li style="flex:1">P<sub style="top: 0.2em;">∞ </sub></li></ul><p></p><p>L<sub style="top: 0.12em;">s </sub>h(z) = </p>
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages28 Page
-
File Size-