Complex measures on path space: An introduction to the Feynman integral applied to the Schrödinger equation (Q1961836)

The Feynman integral is constructed as a genuine integral over a complex measure \(D^{s,t}_x\) on a path space \(\dot\mathbb{R}^{[s,t]}_d\), where \(\dot\mathbb{R}_d\) denotes the one point compactification of the Euclidean space \(\mathbb{R}^d\), and \(\dot\mathbb{R}^{[s,t]}_d\) denotes the infinite product of \([s,t]\) copies of \(\dot\mathbb{R}_d\). The idea is due to \textit{V. P. Maslov} [``Complex Markov chains and the Feynman path integral for nonlinear equations'' (1976; Zbl 0449.35086)], but the exposition in the present paper is more direct and is given in terms of probabilistic concepts generalized to the complex case. The path integral representation of evolutionary equations is given including the Schrödinger equation and dissipative Schrödinger equation. The potential \(V\) in the equations belongs to the space \(F(\mathbb{R}^d)\) of the Fourier transforms of a finite complex measure in \(\mathbb{R}^d\). This assumption is quite restrictive. However, it is not bad from the point of view of calculations, because any function from \(C_0(\mathbb{R}^d)\) can be approximated uniformly by piecewise constant functions, which belong to \(F(\mathbb{R}^d)\). In most cases the complex measure \(D^{s,t}_x\) is absolutely continuous with respect to some real measure of a compound Poisson process, and the corresponding path integral can be computed by means of the Monte Carlo method.