Hamiltonian path integrals in momentum space representation via white noise techniques (Q2450214)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Hamiltonian path integrals in momentum space representation via white noise techniques
scientific article

    Statements

    Hamiltonian path integrals in momentum space representation via white noise techniques (English)
    0 references
    0 references
    19 May 2014
    0 references
    Many attempts have been made to give a meaning to the Hamiltonian path integral (= the Feynman integral in phase space), partly because classical quantum mechanics is based on a Hamiltonian formulation rather than a Lagrangian one, and partly because the semi-classical limit of quantum mechanics is more natural in a Hamiltonian setting, in other words, the phase space is more natural in classical mechanics than the configuration space. In line with the above recognition, this paper provides us with a trial to give a meaning to Hamiltonian integrands in momentum space representation, with the use of a white noise approach (cf. [\textit{T. Hida} et al., White noise. An infinite dimensional calculus. Dordrecht: Kluwer Academic Publishers (1993; Zbl 0771.60048)]). The leading philosophy consists in regarding the Hamiltonian path integrand as a Hida distribution based upon the concepts from \textit{M. Grothaus} and \textit{L. Streit} [Rep. Math. Phys. 44, No. 3, 381--405 (1999; Zbl 1059.81556)]. By means of ambiguity caused by the Heisenberg uncertainty principle in the momentum space representation, the author models the momentum trajectories as a Brownian fluctuation \(B_p\) starting in the initial momentum \(p_0\), just like \[ p(\tau) = p_0 + \frac{ \sqrt{ \hbar m}}{t - t_0} B_p(\tau), \quad 0 \leqslant \tau \leqslant t. \tag{1} \] Furthermore, the space variable is modeled by white noise independent of the Brownian fluctuation, \[ x( \tau) = \sqrt{ \displaystyle{ \frac{ \hbar}{m} } } (t - t_0) \omega_x(\tau), \quad 0 \leqslant \tau \leqslant t. \tag{2} \] Then the author proposes the formal ansatz for the Feynman integrand \(I_V\) in phase space with respect to the Gaussian measure \(\mu\), \[ \begin{aligned} I_V & = N \exp \left\{ \frac{i}{\hbar} \int_{t_0}^t ( - x(\tau) \dot{p}(\tau) - \frac{ p( \tau)^2}{2m} ) d \tau + \frac{1}{2} \int_{t_0}^t ( \omega_x(\tau)^2 + \omega_p(\tau)^2 ) d \tau \right\} \\ & \times \exp \left\{ - \frac{i}{\hbar} \int_{t_0}^t V( x( \tau), p(\tau), \tau) d \tau \right\} \cdot \delta ( p(t) - p' ). \end{aligned}\tag{3} \] In the above expression the sum of the first and the third integral is the action \(S(x, p)\) in momentum representation, and the Donsker delta function \(\delta\) serves to pin trajectories to \(p'\) at time \(t\). The second integral is introduced to simulate the Lebesgue integral by compensation of the fall-off of the Gaussian measure in the time interval \(( t_0, t)\). \(N\) is a normalization which will be implemented by the use of a normalized exponential under the setting of white noise calculus. The construction is done in terms of the \(T\)-transform in white noise analysis, which characterizes Hida distributions. At the same time, the \(T\)-transform of the constructed Feynman integrands provides their generating functional. Then, using the generating functional, the author shows that the generalized expectation (generating functional at zero) gives the Green function to the corresponding Schrödinger equation. Finally, the author proves that the Feynman integrand \(I_{H O, \mathrm{mom}}\) for the harmonic oscilator in phase space in the momentum space exists as a Hida distribution, and also gives an explicit representation of its generating functional. Moreover, its generalized expectation \[ \operatorname{E} [ I_{H O, \mathrm{mom}}] = \sqrt{ \left( \frac{1}{ 2 \pi i \sqrt{k} \sin ( \sqrt{k} t ) } \right) } \exp \left( i \frac{1}{ 2 \sqrt{k} \tan ( \sqrt{k} t) } {p'}^2 \right) \tag{4} \] is the Green function to the Schrödinger equation for the harmonic oscillator in the momentum space. The original idea of realizing Feynman integrals in the framework of white noise analysis stems from a brilliant work by \textit{L. Streit} and \textit{T. Hida} [Stochastic Processes Appl. 16, 55--69 (1984; Zbl 0575.60039)]. For other works on the study of Feynman integrals as a Hida distribution (see, e.g., [\textit{M. de Faria} et al., J. Math. Phys. 32, No. 8, 2123--2027 (1991; Zbl 0754.46050)] and \textit{C. C. Bernido} and \textit{M. V. Carpio-Bernido} [J. Math. Phys. 43, No. 4, 1728--1736 (2002; Zbl 1059.81108)]).
    0 references
    0 references
    0 references
    white noise analysis
    0 references
    Feynman integral
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references