Nonlinear Markov operators, discrete heat flow, and harmonic maps between singular spaces (Q1611261)

The author develops a theory of harmonic maps \(f\colon M\to N\), where \(N\) is a complete metric space of nonpositive curvature in the sense of A. D.\,Alexandrov and \((M,\mathcal M)\) is a measurable space with a given Markov kernel \(p(x,dy)\). For every measurable map \(f\colon M\to N\) and \(x\in M\), let \(Pf(x)\) denote the barycenter of the probability measure \(p(x,f^{-1}(dy))\) on \(N\). The map \(f\) is called \textit{harmonic} on a subset \(D\) of \(M\) if \(Pf=f\) on \(D\). The theory is a nonlinear generalization of the theory of Markov kernels and Markov chains on \(M\). Harmonic maps can be constructed by an explicit nonlinear Markov chain algorithm which in many cases converges exponentially fast. Many smoothing and contraction properties of the linear Markov operator \(P_{M, \mathbb R}\) carry over to the nonlinear Markov operator \(P=P_{M,N}\).