on the correspondence between harmonic analysis and
play

On the correspondence between harmonic analysis and spectral theory - PowerPoint PPT Presentation

On the correspondence between harmonic analysis and spectral theory Zhirayr Avetisyan UCL Maths Imperial, 2017 Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 1 / 24 Outline Physics: elementary particle


  1. On the correspondence between harmonic analysis and spectral theory Zhirayr Avetisyan UCL Maths Imperial, 2017 Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 1 / 24

  2. Outline Physics: elementary particle states Old and new examples General approach Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 2 / 24

  3. Physics: elementary particle states Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 3 / 24

  4. Physics: elementary particle states Quantum system H - complex separable Hilbert space O = { T : H �→ H } - observables, i.e., densely defined normal operators 0 ̸ = C f ⊂ H - a state, i.e., a 1-dimensional Hilbert subspace Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 4 / 24

  5. Physics: elementary particle states Dynamics and symmetries Dynamics: Distinguished observable H ∈ O , time variable t ∈ R , d T dt = ı [ H , T ] , ∀ T ∈ O . Symmetries: A group G of unitary operators U ∈ O such that [ U , H ] = 0 . Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 5 / 24

  6. Physics: elementary particle states Quantum mechanics G - a real Lie group M - a G -manifold ν - a G -invariant measure (volume form) H = L 2 ( M , ν ) U g f ( x ) = f ( g − 1 x ) , ∀ f ∈ H , ∀ g ∈ G H - a G -invariant Laplacian or Schrödinger operator Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 6 / 24

  7. Physics: elementary particle states What is an elementary particle state? Given: S = { T : H → H } - a semigroup of operators Find: (Ω , ˆ ν ) - a measure space ∫ ⊕ F : H → Ω d ˆ ν ( ω ) H ω - a unitary operator (Fourier transform) H ω - irreducible invariant subspace under S for a.e. ω ∈ Ω C f ⊂ H ω - elementary particle states w.r.t. S Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 7 / 24

  8. Physics: elementary particle states Dynamical particles = spectral theory { } H n , n ∈ N S = Ω ∋ ω = ( λ, ω λ ) , λ ∈ σ ( H ) , ω λ ∈ Ω λ ∫ ⊕ ∫ ⊕ F : H → d µ ( λ ) H ω σ ( H ) Ω λ µ - spectral measure, H H ω = λ 1 H ω = C f ω - wave function, spectral mode Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 8 / 24

  9. Physics: elementary particle states Symmetry particles = harmonic analysis/representation theory { } S = U g , g ∈ G G irrep on H ω = H π , π ∈ ˆ Ω ∋ ω = ( π, ω π ) , ω π = 1 , ..., d π ∫ ⊕ d π ⊕ H ω F : H → d ˆ ν ( π ) ˆ G ω π = 1 ν - ’Plancherel’ measure ˆ H ω - Wigner’s elementary particle states Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 9 / 24

  10. Physics: elementary particle states Question : What is an electron, a dynamical or a symmetry particle? Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 10 / 24

  11. Physics: elementary particle states Question : What is an electron, a dynamical or a symmetry particle? Answer : It is both. Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 11 / 24

  12. Old and new examples Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 12 / 24

  13. Old and new examples Example 1: Laplacian on the line. H = ∆ = − ∂ 2 H = L 2 ( R ) . M = R , d ν ( x ) = dx , x , O ∋ X , P, X f ( x ) = xf ( x ) , P = − ı∂ x f ( x ) . ∫ ⊕ ⊕ C e ıλω λ x . Spectral: σ ( H ) = [ 0 , + ∞ ) , H ≃ d λ [ 0 , + ∞ ) ω λ = ± 1 ˆ π ( g ) = e ıπ g . G = R , U g f ( x ) = f ( x − g ) , G = R , H π = C , ∫ ⊕ d π C e ıπ x . Harmonic: H ≃ R Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 13 / 24

  14. Old and new examples Example 2: Laplacian on the plane. M = R 2 , d ν ( x ) = dx 1 dx 2 , H = ∆ = − ∂ 2 x 1 − ∂ 2 H = L 2 ( R 2 ) . x 2 , X i f ( x ) = x i f ( x ) , O ∋ X i , P i , P i = − ı∂ x i f ( x ) , i = 1 , 2. ∫ ⊕ ∫ ⊕ S 1 dS ( ω λ ) C e ı ( λω λ , x ) . Spectral: H ≃ d λλ [ 0 , + ∞ ) G = E ( 2 ) = R 2 ⋊ U ( 1 ) , ˆ gx = e ıϕ x + y , G = R + . g = ( y , ϕ ) , π ( g ) f ( ψ ) = e ıπ y 1 f ( ψ − ϕ ) . H π = L 2 ( S 1 ) , ∫ ⊕ π ( g ) e ı ( λω λ , x ) = e ı ( λω ′ λ , x ) . Harmonic: H ≃ d ππ H π , R + Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 14 / 24

  15. Old and new examples Example 3: Laplacian on the sphere. M = S 2 , d ν ( ϕ ) = sin ϕ 1 d ϕ 1 d ϕ 2 . 1 1 sin ϕ 1 ∂ ϕ 1 sin ϕ 1 ∂ ϕ 1 − sin 2 ϕ 1 ∂ 2 H = L 2 ( S 2 ) . H = ∆ = − ϕ 2 , l ⊕ ∞ ⊕ C Y m σ ( H ) = { l ( l + 1 ) | l ∈ N 0 } , Spectral: H ≃ l ( ϕ ) . l = 0 m = − l ˆ G = SO ( 3 ) , G = N 0 , ⊕ ∞ H π = C { Y m π ( ϕ ) } π Harmonic: H ≃ H π , m = − π . π = 0 Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 15 / 24

  16. Old and new examples Example 4: Solvable Bianchi groups.     1 0 0    e x 3 M  ( x 1 , x 2 , x 3 ) ∈ R 3 g = ( x 1 , x 2 , x 3 ) , x 1 G =  ,  x 2 M ( I ) M ( II ) M ( III ) M ( IV ) M ( V ) ( 0 ) ( 1 ) ( 1 ) ( 1 ) 1 0 1 0 0 0 0 0 0 0 1 0 1 M ( VI q ) , − 1 < q ≤ 1, q ̸ = 0 M ( VII p ) , p ≥ 0 ( 1 ) ( p ) 0 1 0 − q − 1 p Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 16 / 24

  17. Old and new examples Example 4: Solvable Bianchi groups (2). Left and right generators             L 1 1 0 0 ∂ x 1 R 1 0 ∂ x 1  e x 3 M ⊤   =       =    L 2 0 1 0 ∂ x 2 R 2 0 ∂ x 2 ( x 1 , x 2 ) M ⊤ L 3 1 ∂ x 3 R 3 0 0 1 ∂ x 3 h − 1 = h ij R i ⊗ R j Left Haar measure √ det h ∗∗ e − x 3 Tr M dx 1 dx 2 dx 3 . d ν h ( x ) = Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 17 / 24

  18. Old and new examples Example 4: Laplacian on solvable Bianchi groups. M = R 3 , H = L 2 ( R 3 , ν h ) . d ν ( x ) = d ν h ( x ) , H = h ij R i R j + Tr M h 3 i R i , σ ( H ) = [ 0 , + ∞ ) . ∫ ⊕ ∫ ⊕ ν ( k 1 ) dk 2 e k 2 Tr M ⊕ Spectral: H ≃ d λ R 2 d ˆ C ξ k , h ,λ,ω λ ( x ) . [ 0 , + ∞ ) ω λ = ± 1 G ≃ R 2 / e R M ⊤ , ˆ U g f ( x ) = f ( g − 1 x ) , ∫ ⊕ ⊕ ∞ Harmonic: H ≃ d ˆ ν ( π ) H π , U g ξ k , h ,λ,ω λ ( x ) = ξ k ′ , h ,λ,ω λ ( x ) ˆ G ω π = 1 k ′ = e g 3 M ⊤ k , g = ( g 1 , g 2 , g 3 ) ∈ G . Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 18 / 24

  19. General approach Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 19 / 24

  20. General approach Fourier transform G - type I unimodular locally compact group, ν - Haar measure, ˆ G - unitary dual, ˆ ν - Plancherel measure ∫ ⊕ F : L 2 ( G , ν ) → ν ( π ) H π ⊗ H ∗ d ˆ π , ˆ G ∫ ∫ ˆ ν ( π ) Tr [ π ∗ ( x )ˆ f ( π ) = d ν ( x ) f ( x ) π ( x ) , f ( x ) = d ˆ f ( π )] , ˆ G G Plancherel theorem: ∫ ν ( π ) Tr [ˆ f ( π ) ∗ ˆ ∥ f ∥ 2 2 = d ˆ f ( π )] . ˆ G Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 20 / 24

  21. General approach Fourier multipliers U g f ( x ) = f ( g − 1 x ) , f ∈ L 2 ( G , ν ) , g ∈ G . � � U g f ( π ) = π ( g )ˆ f ∗ h ( π ) = ˆ f ( π )ˆ f ( π ) , g ( π ) , where ∫ d ν ( y ) f ( y ) h ( y − 1 x ) . f ∗ h ( x ) = G H ∈ B ( H ) , [ U g , H ] = 0, ∀ g ∈ G , then H f = ˆ � f ( π ) H π , H π ∈ B ( H π ) , i.e., ∫ ⊕ H ≃ d ˆ ν ( π ) × H π . ˆ G Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 21 / 24

  22. General approach Compact groups ˆ G is discrete and d π = dim H π < ∞ , ⊕ L 2 ( G , ν ) ≃ H π ⊗ H ∗ π . ˆ G If { e j } d π j = 1 - eigenfunctions of H π then { } d π ξ π i , j ( x ) = e ∗ i π ∗ ( x ) e j i , j = 1 - eigenfunctions of H, C ( G ) functions. { } ˆ Peter-Weyl theorem: H π ⊗ H ∗ ξ π π = C and i , j { } ˆ π ∈ ˆ ξ π C G dense in C ( G ) . i , j Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 22 / 24

  23. General approach Eigenfunction expansions ’... it is perhaps worth posing the general question of what conditions on ∆ are needed for such a theory to exist (some hints in this direction are in Maurin’66)...’ (R. Strichartz, ’Harmonic Analysis as Spectral Theory of Laplacians’, 1989) Gelfand triple: D ⊂ H ⊂ D ′ , D - nuclear, id : D → H continuous. D - core of H, H : D → D continuous. Eigenfunctions ξ λ,ω λ ∈ D ′ . If H hypoelliptic, then ξ λ,ω λ regular. Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 23 / 24

  24. General approach Decomposition of continuity T π �→ Tr [ T π π ( x )] , L 1 ( H π ) ≃ E π ⊂ C b ( G ) (Godement’52). {∫ } E = d ˆ ν ( π ) α ( π ) , α ( π ) ∈ E π ⊂ C b ( G ) . ˆ G Generalized Peter-Weyl: E is dense in C b ( G ) . Generalized Bochner: � C b ( G ) is the space of E π -valued finite measures on ˆ G . Zhirayr Avetisyan (UCL Maths) Harmonic analysis vs spectral theory Imperial, 2017 24 / 24

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend