Probabilistic behavior of harmonic functions (Q1300350)

The book is devoted to the interplay of potential theory and probability theory. In particular, it is concerned with the behaviour of harmonic functions on the boundary on the one side and martingale theory on the other. The emphasis is led on ``non-tangential behaviour'', i.e., on the limit behaviour of functions in a cone with vertex on the boundary. Troughout, the domain is assumed to be an upper half-space \(\mathbb{R}^{n+ 1}_+\); many of the results are also formulated for ``caloric functions'', i.e. for solutions of the heat equation. The first chapter develops basic ideas, introducing to the ``classical'' potential theory and generalizing the concepts to the non-tangential setup. Thus starting with Poisson kernels, Dirichlet problems and maximal functions on balls and half-spaces, the reader is led to non-tangential versions, \(H^p\)-spaces, singular integrals, (Littlewood-Paley) \(g\)-functions and area-functions for harmonic functions. (Troughout, the formulations and in particular the estimates are to technical to be stated explicitly in a review.) Chapter 2 is concerned with martingale theory -- in particular with ``dyadic martingales'' which are defined by successive decompositions of cubes -- and estimates for maximal functions of martingales. Harmonic functions are decomposed into dyadic martingales and therefore the above-mentioned (probability) estimates may be applied to potential theory, in particular to harmonic functions. Chapter 3 provides in an analogous way -- applying the techniques of Chapter 2 -- the (probabilistic) tools around the LIL (law of iterated logarithm) to obtain sharp estimates in potential theory: starting with Kolmogorov' s LIL for sequences of r.v. the authors present W. Stout's version of the LIL for martingales (outside the set of convergence) which leads to (upper and lower) LIL-estimates for harmonic functions. Chapter 4 presents sharp inequalities for the maximal and area functions \(A\) and \(N\) derived by corresponding estimates for martingales, and Chapter 5 contains estimates for densities of area-integrals. The last chapter finally is concerned with applications in harmonic analysis, in particular with lacunary Fourier series and (sub-)classes of Bloch functions \((F(z)= \sum z^{2^k})\). The reader interested in this subject -- the interplay of probability theory, harmonic analysis and potential theory -- will find a systematic treatment, inspiring both sides, analysis and probability theory.