introduction to control, coherence, and dissipative dynamics · 2011-10-21 · modelization...
TRANSCRIPT
![Page 1: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/1.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Introduction to Control, Coherence, andDissipative Dynamics
Gabriel Turinici
Universite Paris Dauphine and INRIA Rocquencourt 1
IMA, March 1st, 2009
1financial support from IMA, ANR C-QUID, PICS CNRS-NSF and INRIARocquencourt is acknowledged
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 2: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/2.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Single quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
Outline
1 ModelizationSingle quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
2 Observables3 Numerical algorithms
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
4 Inversion and identification approachesIdentifiabilityOptimal identification : implementable experimental +numerical algorithms
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 3: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/3.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Single quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
Single quantum system
Time dependent Schrodinger equation{i ∂∂t Ψ(x , t) = H(t)Ψ(x , t)Ψ(x , t = 0) = Ψ0(x).
(1)
• H(t) = H0+ interaction terms E.g. H0 = −∆ + V (x)• H(t)∗ = H(t) thus ‖Ψ(t)‖L2 = 1, ∀t ≥ 0.• dipole approximation: H(t) = H0 − ε(t)µ(x)• E.g. O − H bond, H0 = − ∆
2m + V , m = reduced mass
V (x) = D0[e−β(x−x0) − 1]2 − D0, µ(x) = µ0xe−x/x∗
• higher order approximation: H(t) = H0 +∑
k εk(t)µk(x)
• misc. approximations (rigid rotor interacting with two-colorlinearly polarized pulse):H(t) = H0 + (E1(t)2 + E2(t)2)µ1 + E1(t)2 · E2(t)µ2
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 4: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/4.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Single quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
Several quantum systems: density matrix formulation
• Motivation (I) : evolution equation for a projector
• Motivation (II) : evolution equation for a sum of projectors
• Equation for the density matrix evolution{i ∂∂t ρ(x , t) = [H(t), ρ(x , t)]ρ(x , t = 0) = ρ0(x).
(2)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 5: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/5.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Single quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
Simultaneous control of quantum systems
Under simultaneous control of a unique laser field L ≥ 2 molecularspecies.
Initial state |Ψ(0)〉 =∏L`=1 |Ψ`(0)〉.
Any molecule evolves by its own Schrodinger equationi~ ∂∂t |Ψ`(t)〉 = [H`
0 − µ` · ε(t)]|Ψ`(t)〉
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 6: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/6.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Single quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
A set of identical molecules with different spatial positions
N ≥ 2 identical molecules, DIFFERENT orientations, simultaneouscontrol by one laser field.Interaction with a field µ(x)ε(t)u(R) where R = (r , θ, ζ)characterizes the localization of the molecule in the ensemble.
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 7: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/7.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Single quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
Evolution semigroup
{i ∂∂t U(t) = H(t)U(t)U(0) = Id .
(3)
Relationship with wavefunction and density matrix versions: ...
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 8: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/8.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Outline
1 ModelizationSingle quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
2 Observables3 Numerical algorithms
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
4 Inversion and identification approachesIdentifiabilityOptimal identification : implementable experimental +numerical algorithms
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 9: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/9.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Observables: localization
e.g. O-H bond : O(x) = γ0√π
e−γ20 (x−x ′)2
Figure: Successful quantum control for the localization observable.
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 10: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/10.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Observables: projections on eigenstates
Other important example: projection on eigenstatesdensity matrix formulation Tr(ρO)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 11: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/11.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Outline
1 ModelizationSingle quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
2 Observables3 Numerical algorithms
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
4 Inversion and identification approachesIdentifiabilityOptimal identification : implementable experimental +numerical algorithms
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 12: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/12.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Galerkin discretization of the Time Dependent Schrodingerequation
i∂
∂tΨ(x , t) = (H0 − ε(t)µ)Ψ(x , t)
• basis functions {ψi ; i = 1, ...,N}, e.g. the eigenfunctions of theH0: H0ψk = ekψk
• wavefunction written as Ψ =∑N
k=1 ckψk
• construct the matrices (N × N) associated to the operators H0
and µ : Akl = 〈ψk |H0|ψl〉, Bkl = 〈ψk |µ|ψl〉,
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 13: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/13.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Galerkin discretization of the Time Dependent Schrodingerequation
• Finite dimensional equations{i~ ∂∂t c = (A− ε(t)B)c
c(t = 0) = c0
The system evolves on the complex unit sphere SN−1C of
CN :∑N
i=1 |ci |2(t) = 1, ∀t ≥ 0.
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 14: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/14.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Simultaneous control of quantum systems: discretization
Discretization spaces D` = {ψ`i (x); i = 1, ..,N`}, N`,N` ≥ 3eigenstates of H`
0
A` and B` = matrices of the operators H`0 and µ` respectively,
with respect to D`; N =∑L
`=1 N`,
A =
A1 0 . . . 00 A2 . . . 0...
.... . .
...0 0 . . . AL
, B =
B1 0 . . . 00 B2 . . . 0...
.... . .
...0 0 . . . BL
.
Note: the system evolves on the product of spheresS =
∏L`=1 S
N`−1C , Sk−1
C = complex unit sphere of Ck .
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 15: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/15.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Figure: Polarization-shaped pulse, optimized for the ionization ofpotassium molecules. Ellipses represent the amplitude of the electricfield, colours indicate different frequencies; Yaron Silberberg, Nature 430,624-625 (2004)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 16: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/16.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Monotonic algorithms
• evaluation of the quality of a control through a objectivefunctional to maximizeJ(ε) = 2<〈ψtarget |ψ(·,T )〉 −
∫ T0 α(t)ε2(t)dt
J(ε) = 2− ‖ψtarget − ψ(·,T )‖2L2 −
∫ T0 α(t)ε2(t)dt
J(ε) = 〈Ψ(T )|O|Ψ(T )〉 − α∫ T
0 ε2(t)dt
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 17: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/17.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Standard optimization procedure
• construction of an extended objective functional i.e., addconstraints through an adjoint state χ(x , t)
J(ε) = 〈Ψ(T )|O|Ψ(T )〉 − α∫ T
0ε2(t)dt
−2Re
∫ T
0
⟨χ(x , t),
{∂
∂t+ i · [H0 − ε(t)µ]
}Ψ(x , t)
⟩Partial derivatives
δJ(ε)
δε= −2αε(t)− 2Im 〈χ|µ|Ψ〉 (t)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 18: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/18.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Euler-Lagrange critical point equation
{i ∂∂t Ψ(x , t) = (H0 − ε(t)µ)Ψ(x , t)Ψ(x , t = 0) = Ψ0(x){i ∂∂tχ(x , t) = (H0 − ε(t)µ)χ(x , t)χ(x , t = T ) = OΨ(x ,T )
αε(t) = −Im 〈χ|µ|Ψ〉 (t)
• Chose a numerical algorithm to update the field ε(t), e.g.,
εn+1 = εn +δJ(εn)
δε(4)
slow convergence =⇒ complicated objective functional surfaceRecent works by Alfio Borzi: functional surface seems to be veryflat with many almost optimal regions.
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 19: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/19.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Compute the optimal field ε(t) (Krotov cf. Tannor et. al 1992):(χk−1, εk−1,Ψk−1)→ (χk , εk ,Ψk){
i ∂∂t Ψk(x , t) = (H0 − εk(t)µ)Ψk(x , t)Ψk(x , t = 0) = Ψ0(x)
(5)
εk(t) = − 1
αIm〈χk−1|µ|Ψk〉(t) (6){
i ∂∂tχk(x , t) = (H0 − εk(t)µ)χk(x , t)
χk(x , t = T ) = OΨk(x ,T )(7)
In practice solve the equations (5)-(6) by propagating thenon-linear equation{
i ∂∂t Ψk(x , t) = (H0 + 1α Im〈χk−1|µ|Ψk〉(t)µ)Ψk(x , t)
Ψk(x , t = 0) = Ψ0(x)(8)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 20: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/20.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Zhu & Rabitz formulation (1998)
{i ∂∂t Ψk(x , t) = (H0 − εk(t)µ)Ψk(x , t)Ψk(x , t = 0) = Ψ0(x)
εk(t) = − 1
αIm〈χk−1|µ|Ψk〉(t){
i ∂∂tχk(x , t) = (H0 − εk(t)µ)χk(x , t)
χk(x , t = T ) = OΨk(x ,T )
εk(t) = − 1
αIm〈χk |µ|Ψk〉(t)
THEOREM (W. Zhu and H. Rabitz. J. Chem. Phys., 109:385–391,
1998.) Suppose O is a semi-positive definite (auto-adjoint)operator. Then for any k ≥ 0: J(εk+1) ≥ J(εk), i.e. there is animprovement in the functional at any iteration.
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 21: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/21.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 22: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/22.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Other classes of algorithms (Y.Maday & G.T. 2002), otherworks by Y. Ohtsuki, S. Schirmer, D. Sugny, etc
{i ∂∂t Ψk(x , t) = (H0 − εk(t)µ)Ψk(x , t)Ψk(x , t = 0) = Ψ0(x)
(9)
εk(t) = (1− δ)εk−1(t)− δ
αIm〈χk−1|µ|Ψk〉(t) (10){
i ∂∂tχk(x , t) = (H0 − εk(t)µ)χk(x , t)
χk(x , t = T ) = OΨk(x ,T )(11)
εk(t) = (1− η)εk(t)− η
αIm〈χk |µ|Ψk〉(t) (12)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 23: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/23.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
THEOREM If O is an hermitian observable semi-positive definitethen, for any η, δ ∈ [0, 2] J(εk+1) ≥ J(εk).
J(εk+1)− J(εk) =⟨Ψk+1(T )−Ψk(T )|O|Ψk+1(T )−Ψk(T )
⟩+
α
∫ T
0(
2
δ− 1)(εk+1 − εk)2 + (
2
η− 1)(εk − εk)2
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 24: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/24.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Experimental control
Figure: Experimental quantum control: In practice H and/or the dipole may not be known: one uses then azero order algorithm (no gradients): genetic algorithms, Nelder-Mead simplex (C. Le Bris, H.Rabitz & G.T. PRE
2004) ... This is possible due to a high laboratory experimental repetition rate (1Hz − 105Hz or more).
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 25: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/25.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
Lyapounov (tracking) approaches
keep decreasing ‖ψ(t)− φ0‖2 ...
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 26: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/26.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
IdentifiabilityOptimal identification : implementable experimental + numerical algorithms
Outline
1 ModelizationSingle quantum system: wavefunction formulationSeveral quantum systems: density matrix formulationSimultaneous control of quantum systemsEvolution semigroup
2 Observables3 Numerical algorithms
Simultaneous control of quantum systems: discretizationMonotonically convergent algorithmsExperimental controlLyapounov (tracking) approaches
4 Inversion and identification approachesIdentifiabilityOptimal identification : implementable experimental +numerical algorithms
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 27: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/27.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
IdentifiabilityOptimal identification : implementable experimental + numerical algorithms
Inversion and identification approaches
here µ and/or H0 and/or ψ(0) are unknown, need to recovedfrom measurements
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 28: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/28.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
IdentifiabilityOptimal identification : implementable experimental + numerical algorithms
Identifiability
Theorem (C. Le Bris, M. Mirrahimi H. Rabitz, G.T.; 2006 )Consider two finite dimensional Hamiltonians H1, H2 and twodipole moments µ1, µ2, with Ψ1,Ψ2 ,solutions of :
iΨk = (Hk + ε(t) µk) Ψk
Suppose ∀t ≥ 0, ∀ε(t) ∈ L2
|〈Ψ1(t) | ei 〉|2 = |〈Ψ2(t) | ei 〉|2 i = 1, ..,N, (13)
( {ei}Ni=1 = canonical basis of CN).
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 29: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/29.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
IdentifiabilityOptimal identification : implementable experimental + numerical algorithms
Assume for H1, H2 :
1 Equations are controllable
2 H1 and H2 have the same eigenvalues λi .
3 λi1 − λj1 6= λi2 − λj2 for (i1, j1) 6= (i2, j2).
4 〈φi | µ | φi 〉 = 0, i = 1, ..,N.
5 there does not exist a subspace of dimension one or twospanned by the vectors {ei}, which remains invariant duringthe free evolution (ε ≡ 0) of the first system (H1 and µ1).
Then, there exist {αi}Ni=1 such that, for all 1 ≤ i , j ≤ N,
(µ1)ij = e i(αi−αj )(µ2)ij , (H1)ij = e i(αi−αj )(H2)ij . (14)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 30: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/30.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
IdentifiabilityOptimal identification : implementable experimental + numerical algorithms
Optimal identification implementable experimental + numerical algorithms
1 Objective functional J(ε, µ,V ) the distance between the measures with the field ε(t)
and the numerical simulation with the potential V , dipole µ and field ε.“Canonical” formulation of the inversion problem : solve
minV ,µ
∫L2
J(ε, µ,V )dε(t) (15)
BUT: space L2 too large ... introduce a discriminating fieldapproach that allows to distinguish between the admissiblecandidates.
2 Inversion problem: S(ε) = {µ,V ; J(ε, µ,V ) ≤ tolerance}.m(ε) = a measure of S(ε).
3 Optimization problem: minimize m(ε). The solution ε is thediscriminating field; S(ε) = set of possible solutionsIn practice: m(ε)= diameter, S(ε) not explicitly computed (!).Algorithms can use fields coming from the identifiability theory.
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics
![Page 31: Introduction to Control, Coherence, and Dissipative Dynamics · 2011-10-21 · Modelization Observables Numerical algorithms Inversion and identi cation approaches Introduction to](https://reader035.vdocuments.mx/reader035/viewer/2022081613/5fb5d8e3eca6bd35bb36afd5/html5/thumbnails/31.jpg)
ModelizationObservables
Numerical algorithmsInversion and identification approaches
IdentifiabilityOptimal identification : implementable experimental + numerical algorithms
Figure: Optimal identification machine : implementable experimental +numerical algorithms (J.M. Geremia & H. Rabitz JCP, 2003)
Gabriel Turinici Introduction to Control, Coherence, and Dissipative Dynamics