Dispersion and moment lemmas revisited (Q1303842)

From MaRDI portal
Revision as of 21:55, 30 July 2024 by Daniel (talk | contribs) (‎Created claim: Wikidata QID (P12): Q124811198, #quickstatements; #temporary_batch_1722372779079)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Dispersion and moment lemmas revisited
scientific article

    Statements

    Dispersion and moment lemmas revisited (English)
    0 references
    25 October 1999
    0 references
    The authors' opening sentences state: ``We investigate relations between the dispersive effects for various classical equations (Schrödinger, Dirac, and wave equation). After Wigner transform dispersive effects are reduced to moment lemmas for kinetic equations yielding new results.'' The authors first analyze solutions of the transport equations of the type: \(f_t+\xi \cdot \nabla_xf=g\), \(\{x,\xi\}\in \mathbb{R}^{2n}\), \(t\in\mathbb{R}\), with initial non-negative value of \(f\). They introduce the multiplier, depending on parameters \(\alpha, \sigma,\delta\), producing a time integrated \(x\)-weighted \(\xi\)-moment of \(f\) of the order \(2+\sigma\). The authors offer several examples of application. For the Liouville equation they insert \(\delta=0\), and \(x_0= 0\), into the multiplier formula \(\mu=|\xi|^\sigma(x-x_0)\cdot\xi/ \{( \delta+ |x-x_0|^\alpha)\}^{1/\alpha}\), \(\alpha>0\), \(\delta>0\), \(\sigma \in\mathbb{R}\). The resulting formula establishes a relation between the \(\xi\)-moment of \(f\) of the order \(2+\sigma\), and the moment of order \(1+\sigma\), (except in the \(x=\xi\) direction). They prove several crucial inequalities using this formulation. They also prove some energy estimates for the nonlinear Vlasov equation, and note similarity of these estimates with the Morawetz estimate for the nonlinear Schrödinger equation. They also show that their technique works for the transport equation \(f_t+\underline v(\xi)\cdot \nabla_xf=0\), and finally prove a trace formula that can be obtained by putting \(\mu= \text{sign}_\varepsilon (x\cdot\xi)\). The next section called ``Dispersion Lemmas for Schrödinger-type problems'' uses the Wigner transform. Basically it transforms \(A(x,\nu)\in S'(\mathbb{R}^n,\mathbb{C})\) into \(w^\varepsilon[f] (x,\xi)\) by taking Fourier transform of \(A\overline A\) (the conjugate transpose of \(A)\), with \(x\) perturbed by \(1/2\varepsilon\nu\) in \(\overline A\) and by \(-1/2\varepsilon\nu\) in \(A\). Here \(\xi\) pretends to be a frequency-like vector, while perturbation \(\nu\) pretends to be a vector of space or time-like variables, if we try to form an electrical engineering analogy. The translation by \(v\) inside the Fourier transform has an analogy in the wavelet transform, which was of course invented many years after Eugene Wigner considered this class of problems. Several other problems such as the nonlinear Schrödinger equation, the equations of quantum hydrodynamics, are given the same treatment. The last section introduces radial multipliers allowing the authors to establish some properties of the free Dirac equation in \(\mathbb{R}^3\) on the spinor field \((\mathbb{C}^4)\).
    0 references
    0 references
    Schrödinger and Dirac equations
    0 references
    Wigner transform
    0 references
    dispersive estimates
    0 references
    kinetic equations
    0 references
    transport equations
    0 references
    Liouville equation
    0 references
    energy estimates
    0 references
    nonlinear Vlasov equation
    0 references
    nonlinear Schrödinger equation
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references