English

Nearly hyperharmonic functions are infima of excessive functions

Probability 2019-06-06 v3

Abstract

Let X\mathfrak X be a Hunt process on a locally compact space XX such that the set EX\mathcal E_{\mathfrak X} of its Borel measurable excessive functions separates points, every function in EX\mathcal E_{\mathfrak X} is the supremum of its continuous minorants in EX\mathcal E_{\mathfrak X} and there are strictly positive continuous functions v,wEXv,w\in\mathcal E_{\mathfrak X} such that v/wv/w vanishes at infinity. A numerical function u0u\ge 0 on XX is said to be nearly hyperharmonic, if uXτVdPxu(x)\int^\ast u\circ X_{\tau_V}\,dP^x\le u(x) for all xXx\in X and relatively compact open neighborhoods VV of xx, where τV\tau_V denotes the exit time of VV. For every such function uu, its lower semicontinous regularization u^\hat u is excessive. The main purpose of the paper is to give a short, complete and understandable proof for the statement that every Borel measurable nearly hyperharmonic function on XX is the infimum of its majorants in EXE_{\mathfrak X}. The major novelties of our approach are the following: 1. A quick reduction to the special case, where starting at xXx\in X with u(x)<u(x)<\infty the expected number of times the process X\mathfrak X visits the set of points yXy\in X, where u^(y):=lim infzyu(z)<u(y)\hat u(y):=\liminf_{z\to y} u(z)<u(y), is finite. 2. The statement that the integral udμ\int u\,d\mu is the infimum of all integrals wdμ\int w\,d\mu, wEXw\in E_{\mathfrak X} and wuw\ge u, not only for measures μ\mu satisfying wdμ<\int w\,d\mu<\infty for some excessive majorant ww of uu, but also for all finite measures. At the end, the measurability assumption on uu is weakened considerably.

Keywords

Cite

@article{arxiv.1809.08611,
  title  = {Nearly hyperharmonic functions are infima of excessive functions},
  author = {Wolfhard Hansen and Ivan Netuka},
  journal= {arXiv preprint arXiv:1809.08611},
  year   = {2019}
}

Comments

The presentation is improved at various places. In particular, the special case is more restrictive and yields a better intuition, there is a new Lemma 3.5 leading to a simplification in the proof of Theorem 3.4, and the reduction to the special case in Section 4 is shortened. Whereas Sections 5 and 6 are not modified, there is a more general Section 6

R2 v1 2026-06-23T04:15:24.052Z