research33 min studyLesson 826 in Mathematics

Egorov's theorem and classical transport

Transport observables with a sign-consistent Hamiltonian flow and state the fixed-time scope before considering logarithmically long times.

On this page

What you’ll understand

  • Match Heisenberg evolution to pullback by the Hamiltonian flow.
  • State fixed-time operator Egorov under controlled symbol and flow hypotheses.
  • Verify exact quadratic transport and explain the conditional nature of Ehrenfest-time extensions.

A quantum state evolves in Hilbert space, while a classical observable evolves by composition with a phase-space flow. Egorov's theorem connects these two descriptions for suitable symbols and time scales. Its direction depends on whether an observable is conjugated by U*AU or UAU*, and its error depends on derivatives of the transported symbol. Exact quadratic examples isolate the sign and algebra before a general approximation theorem is used.

Heisenberg evolution fixes the direction

Let P_h be self-adjoint and set U_h(t)=exp(−itP_h/h). The Heisenberg observable is U_h(t)* A U_h(t), whose derivative is (i/h) times its commutator with P_h. Under our bracket convention, the leading commutator symbol is (h/i){p,a}; the two coefficients cancel, leaving H_p a={p,a}. Thus the leading symbol is a composed with the forward Hamiltonian flow Φ_t. Conjugating in the opposite order reverses that flow direction.

A concrete sufficient setting is the mass-one Schrödinger operator with smooth real subquadratic potential, meaning derivatives of V of order at least two are bounded, with its self-adjoint realization. The classical flow is complete and has controlled derivatives on every fixed time interval. Start with an h-independent compact smooth phase-space observable and follow its compact flow tube for |t|≤T. These support and derivative controls are what make the operator remainder uniform on that specified interval.

Uh(t)=e−itPh/h,ddt(Uh(t)∗AUh(t))=ih[Ph,Uh(t)∗AUh(t)],∂t(a∘Φt)={p,a∘Φt}U_h(t)=e^{-itP_h/h},\qquad\frac d{dt}(U_h(t)^*AU_h(t))=\frac i h[P_h,U_h(t)^*AU_h(t)],\qquad\partial_t(a\circ\Phi_t)=\{p,a\circ\Phi_t\}
The forward-flow symbol corresponds specifically to U*AU with the declared Poisson bracket and Fourier signs.↔ Scroll wide equations horizontally.

A fixed-time theorem is not an all-time limit

In this setting, the leading Egorov statement gives U* a^w U=(a composed with Φ_t)^w+O(h) in L² operator norm uniformly for |t|≤T, with T fixed independently of h. Higher symbol expansions can improve the remainder when their hypotheses and subprincipal terms are tracked. The leading statement does not identify the full propagated operator exactly, and the constant can grow when the time interval grows. Compact support at the initial time alone is not a bound on every derivative at arbitrarily long times.

Long-time results require additional quantitative flow control. If a particular proof yields an error bounded by C h exp(Γ|t|), substituting |t|≤c log(1/h) gives C h^(1−Γc), which tends to zero only when c<1/Γ. This conditional calculation explains an Ehrenfest-type logarithmic scale; it is not a universal theorem with a universal c for every Hamiltonian. Hyperbolic derivative growth and the symbol class's allowed derivative losses determine which long-time statement can actually be proved.

Uh(t)∗awUh(t)=(a∘Φt)w+OL2→L2(h)(∣t∣≤T),heΓclog⁡(1/h)=h1−ΓcU_h(t)^*a^wU_h(t)=(a\circ\Phi_t)^w+O_{L^2\to L^2}(h)\quad(|t|\le T),\qquad h e^{\Gamma c\log(1/h)}=h^{1-\Gamma c}
The first estimate has fixed T. The second identity explains the consequence of an additionally established exponential-in-time error estimate.↔ Scroll wide equations horizontally.

Quadratic dynamics provide an exact calibration

For a quadratic Hamiltonian, the Weyl commutator closes exactly because the higher odd derivative terms vanish. Linear and quadratic observables therefore evolve by exact Heisenberg identities on a common invariant Schwartz domain. These unbounded observables are domain identities, not examples of a bounded operator-norm remainder without cutoffs. For compact observables, metaplectic covariance supplies the corresponding exact quadratic Egorov relation. This exceptional exactness does not extend unchanged to a general nonlinear potential.

The same theorem transports limiting measures: testing an evolved state against a compact observable gives the initial measure tested against its composition with Φ_t, hence a pushforward by the forward flow. The result is for a fixed time and an identified subsequence, with localization preserving the required mass control. It describes quadratic expectations rather than a claim that the wavefunction becomes one classical trajectory. A coherent packet can spread while its leading phase-space center follows the classical motion.

Free motion moves Q forward

For P_h=P²/2 with P=−ih∂x and Q=x, compute U*QU and U*PU. Compare the answer with the classical free flow.

  1. The canonical commutator gives [P²/2,Q]=−ihP. Multiplying by i/h yields the Heisenberg derivative P, while [P²/2,P]=0 keeps momentum constant.

  2. Integrating with initial values Q and P gives Q(t)=Q+tP and P(t)=P on Schwartz functions. These identities are exact and their plus sign agrees with forward motion.

  3. The classical symbol p=ξ²/2 has equations xdot=ξ and ξdot=0, so Φ_t(x,ξ)=(x+tξ,ξ). Pulling the observable x back by that flow gives x+tξ, whose Weyl quantization is Q+tP.

The result

The exact Heisenberg and classical formulas agree: Q(t)=Q+tP. Reversing the conjugation would reverse the sign of t.

Explore the model

Interactive lab

Interactive lab · oscillator

An exactly solvable quadratic flow

Pause and Reset, then set damping ζ = 0 and frequency ω₀ = 1. After two Steps, t=1: predict x=cos1≈0.540302 and momentum equal to the displayed velocity −sin1≈−0.841471. Reset and set both ζ = 0 and ω₀ = 2, then take two Steps: x=cos2 and momentum −2sin2. These are classical flows for (ξ²+ω₀²x²)/2 from (1,0). The instrument does not evolve a wavefunction or measure an Egorov remainder.

An exactly solvable quadratic flowUnit mass, x(0) = 1 and v(0) = 0. The analytic solution includes underdamped, critically damped and overdamped cases; time is measured in seconds. Current readings: Displacement x(t): 1; Velocity v(t): 0; Regime: underdamped. All parameters are adjustable with the labeled controls below.02.557.510−1.15−0.5700.571.15Time · sDisplacementunderdamped
Displacement x(t)Current time
Displacement x(t)
1
Velocity v(t)
0
Regime
underdamped
t = 0 s

Unit mass, x(0) = 1 and v(0) = 0. The analytic solution includes underdamped, critically damped and overdamped cases; time is measured in seconds. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forQuadratic flow is an exact sign check; general fixed-time approximations still require symbol and derivative estimates.

Check your understanding

Which leading symbol corresponds to U(t)*aʷU(t) when U(t)=exp(−itP/h) and H_p a={p,a}?

Try it yourself

A coherent state has initial means E Q=1/2, E P=1, variances h/2 in both coordinates, and zero symmetrized covariance. Under free evolution, compute E Q(2), Var(Q(2)), and E Q(2)².

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

Use the exact Q(2)=Q+2P identity and expand its square with the symmetrized mixed term, rather than treating QP as a real scalar product.

Compare with the solution

The mean is 1/2+2=5/2. The variance is h/2+4(h/2)=5h/2 because the symmetrized covariance is zero. Thus EQ(2)2=25/4+5h/2\mathbb E Q(2)^2=25/4+5h/2. The operator square is Q²+2(QP+PQ)+4P², the Weyl quantization of (x+2ξ)². This domain calculation is exact for the coherent Schwartz state; it is not an application of the compact-observable operator-norm theorem to an arbitrary unbounded polynomial without specifying a state and domain.

Take it with you

  • Heisenberg U*AU transports the leading symbol along the forward classical flow under the fixed conventions.
  • Fixed-time errors do not automatically remain small on h-dependent time intervals.
  • Quadratic exactness checks the formulas but is exceptional among general Hamiltonians.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.