2016/10/13 by Gerhard Keller, Keller, Gerhard
Mathematics · #37A30 #37B05 #Dynamical Systems (math.DS) #FOS: Mathematics #math.DS #msc:37A30 #msc:37B05
paper · pdf · doi:10.48550/arxiv.1610.03998
v2: A minor remark on the bottom of page 2 was corected. v3: The statement of Theorem 2.4 was corrected and a missing piece of the proof was provided. v4: This version combines a modification of v3 (in Section 3) with some new material in Section 4
arxiv created 2019/05/15 · arxiv updated 2019/05/16
The orbit closures of regular model sets generated from a cut-and-project scheme given by a co-compact lattice L⊂ G× H and compact and aperiodic window W⊆ H, have the maximal equicontinuous factor (MEF) (G× H)/L, if the window is toplogically regular. This picture breaks down completely, when the window has empty interior, in which case the MEF is always trivial, although (G× H)/L continues to be the Kronecker factor for the Mirsky measure. As this situation occurs for many interesting examples like the square-free numbers or the visible lattice points, there is some need for a slightly weaker concept of topological factors that is still strong enough to capture basic properties of the system. Here we propose to use the concept of a generic factor \citeHuangYe2012 for this purpose. For so called ergodic topological dynamical systems we prove the existence of a maximal equicontinuous generic factor (MEGF) and characterize it in terms of the regional proximal relation. For such systems we also show that the MEGF is trivial if and only if the system is topologically weakly mixing. This part of the paper profits strongly from previous work by McMahon \citeMcMahon1978 and Auslander \citeAuslander1988. In Section 3 we show that (G× H)/L is indeed the MEGF of the orbit closure of each weak model set with an aperiodic Haar regulaar window, and in Section 4 we apply this fact to give an alternative proof of the finding \citeMentzen2017 that the centralizer of any B-free dynamical system of Erdös type is trivial.