Kramers problem: Numerical Wiener-Hopf-like model characteristics

Since the Kramers problem cannot be, in general, solved in terms of elementary functions, various numerical techniques or approximate methods must be employed. We present a study of characteristics for a particle in a damped well, which can be considered as a discretized version of the Melnikov [Phys. Rev. E 48, 3271 (1993)] turnover theory. The main goal is to justify the direct computational scheme to the basic Wiener-Hopf model. In contrast to the Melnikov approach, which implements factorization through a Cauchy-theorem-based formulation, we employ the Wiener-Levy theorem to reduce the Kramers problem to a Wiener-Hopf sum equation written in terms of Toeplitz matrices. This latter can provide a stringent test for the reliability of analytic approximations for energy distribution functions occurring in the Kramers problems at arbitrary damping. For certain conditions, the simulated characteristics are compared well with those determined using the conventional Fourier-integral formulas, but sometimes may differ slightly depending on the value of a dissipation parameter. Another important feature is that, with our method, we can avoid some complications inherent to the Melnikov method. The calculational technique reported in the present paper may gain particular importance in situations where the energy losses of the particle to the bath are a complex-shaped function of the particle energy and analytic solutions of desired accuracy are not at hand. In order to appreciate more readily the significance and scope of the present numerical approach, we also discuss concrete aspects relating to the field of superionic conductors.