Affine maps between CAT(0) spaces (Q2634841): Difference between revisions

The authors study affine maps between \(\mathrm{CAT}(0)\) spaces admitting cocompact isometric actions. A map between geodesic spaces \(f: X\rightarrow Y\) is affine if it maps geodesics \(\gamma\) in \(X\) to geodesics in \(Y\) and rescale \(\gamma\) with constant speed \(\rho(\gamma)\), which may depend on the geodesic \(\gamma\). The main result of the paper says that if \(X\) is a geodesically complete proper \(\mathrm{CAT}(0)\) space admitting a cocompact isometric action and \(f: X\rightarrow Y\) is a continuous affine map into a \(\mathrm{CAT}(0)\) space, then \(f\) splits as a product of dilations (on non-Euclidean factors of \(X\)) and a linear map (on the Euclidean factor of \(X\)). A key tool in the proof is a splitting lemma for the Tits boundary of \(\mathrm{CAT}(0)\) spaces. This lemma gives a sufficient condition for the Tits boundary to split as a spherical join. A subset \(P\) of the Tits boundary of a \(\mathrm{CAT}(0)\) space \(X\) is called \(\pi\)-convex if for any \(x, y\in P\) with Tits distance \(d_T(x,y)<\pi\) the unique Tits geodesic from \(x\) to \(y\) lies in \(P\). A non-empty subset \(P\) of the Tits boundary is involutive if \(x\in P\) implies any that \(y\) in the Tits boundary satisfying \(d_T(x, y)\geq \pi\) also lies in \(P\). The splitting lemma says that if \(X\) is a proper \(\mathrm{CAT}(0)\) space with cocompact isometric group action and \(P\) is a non-empty subset of the Tits boundary that is \(\pi\)-convex and involutive, then the Tits boundary splits as a spherical join with \(P\) as a factor.