spectral projections and resolvent bounds for quantized...

5
Spectral projections and resolvent bounds for quantized partially elliptic quadratic forms Joe Viola Laboratoire de Mathématiques Jean Leray Université de Nantes [email protected] 14 March 2013

Upload: others

Post on 21-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

  • Spectral projections and resolvent bounds forquantized partially elliptic quadratic forms

    Joe Viola

    Laboratoire de Mathématiques Jean LerayUniversité de Nantes

    [email protected]

    14 March 2013

  • Examples and resolvent growth

    We will discuss the class of elliptic(ish) quadratic operators Q.As examples, consider the complex harmonic oscillator

    Qθ = −e−2iθd2

    dx2+ e2iθx2,

    studied by E. B. Davies and others, and quadraticKramers-Fokker-Planck

    Pa =12

    (v2 − ∂2v ) + a(v∂x − x∂v).

    Question: What is the behavior of the resolvent norm

    ||(Q− z)−1||L(L2), |z| → ∞?

    Particularly, what is the rate of exponential growth deep withinthe numerical range, close to the spectrum?

  • log10 ||(Pa − z)−1||L(L2) for a = 1/√

    2

    0 5 10 150

    2

    4

    6

    8

    10

    12

    14

    16

    18

    0

    2

    4

    6

    8

    10

    12

    14

  • Spectral projections

    We may define spectral projections around an eigenvalue “µα”via

    Πα =1

    2πi

    ∫|z−µα|=ε

    (z− Q)−1 dz.

    If ||Πα|| large, then ||(Q− z)−1|| somewhere large, but not viceversa (cancellation).

    We may find vα with ||vα|| = 1 and

    ||Πα|| = ||Παvα||.

    If ||Παvα|| is large, maybe so is ||(Q− z)−1vα||?

  • Good approximation by ||(Pa − z)−1vα|| for a = 1/√

    2For corresponding α, compute relative error

    log10

    (||(Pa − z)−1||||(Pa − z)−1vα||

    − 1).

    10.5 11 11.5

    2.8

    3

    3.2

    3.4

    3.6

    3.8

    4

    4.2

    10.5 11 11.5

    2.8

    3

    3.2

    3.4

    3.6

    3.8

    4

    4.2

    10.5 11 11.5

    2.8

    3

    3.2

    3.4

    3.6

    3.8

    4

    4.2

    10.5 11 11.5

    2.8

    3

    3.2

    3.4

    3.6

    3.8

    4

    4.2

    10.5 11 11.5

    2.8

    3

    3.2

    3.4

    3.6

    3.8

    4

    4.2

    7

    6

    5

    4

    3

    2

    1

    0

    1

    2

    3