Averaging of random walks and shift-invariant measures on a Hilbert space
We study random walks in a Hilbert space H and representations using them of solutions of the Cauchy problem for differential equations whose initial conditions are numerical functions on H. We construct a finitely additive analogue of the Lebesgue measure: a nonnegative finitely additive measure λ that is defined on a minimal subset ring of an infinite-dimensional Hilbert space H containing all infinite-dimensional rectangles with absolutely converging products of the side lengths and is invariant under shifts and rotations in H. We define the Hilbert space H of equivalence classes of complex-valued functions on H that are square integrable with respect to a shift-invariant measure λ. Using averaging of the shift operator in H over random vectors in H with a distribution given by a one-parameter semigroup (with respect to convolution) of Gaussian measures on H, we define a one-parameter semigroup of contracting self-adjoint transformations on H, whose generator is called the diffusion operator. We obtain a representation of solutions of the Cauchy problem for the Schrödinger equation whose Hamiltonian is the diffusion operator.
Keywordsinvariant measure on Hilbert space finitely additive measure random walk Schrödinger equation Cauchy problem
Unable to display preview. Download preview PDF.
- 1.A. D. Wentzell and M. I. Freidlin, Fluctuations in Dynamical Systems Caused by Small Random Pertubations [in Russian], Nauka, Moscow (1979).Google Scholar
- 3.I. Daletskii and S. V. Fomin, Measures and Differential Equations in Infinite-Dimensional Spaces [in Russian], Nauka, Moscow (1983).Google Scholar
- 22.L. A. Borisov, Yu. N. Orlov, and V. Zh. Sakbaev, “Feynman formulas for averaging of semigroups generated by operators of Schrödinger type [in Russian],” Preprint No. 057, Keldysh Inst. of Applied Math., Moscow (2015).Google Scholar
- 24.O. G. Smolyanov and E. T. Shavgulidze, Continual Integrals [in Russian], URSS, Moscow (2015).Google Scholar