An Ultrafilter Approach to Jin’s Theorem
نویسنده
چکیده
It is well known and not difficult to prove that if C ⊆ Z has positive upper Banach density, the set of differences C −C is syndetic, i.e. the length of gaps is uniformly bounded. More surprisingly, Renling Jin showed that whenever A and B have positive upper Banach density, then A − B is piecewise syndetic. Jin’s result follows trivially from the first statement provided that B has large intersection with a shifted copy A−n of A. Of course this will not happen in general if we consider shifts by integers, but the idea can be put to work if we allow “shifts by ultrafilters”. As a consequence we obtain Jin’s Theorem. The upper Banach density of C ⊆ Z is given by d(C) := limm−n→∞ 1 m−n+1 |C∩{n, . . . ,m}|. A set S ⊆ Z is syndetic if the gaps of S are of uniformly bounded length, i.e. if there exists some k > 0 such that S −{−k, . . . , k} = Z. A set P ⊆ Z is piecewise syndetic if it is syndetic on large pieces, i.e. if there exists some k ≥ 0 such that P − {−k, . . . , k} contains arbitrarily long intervals of integers. It was first noted by Følner ([Føl54a, Føl54b]) that C − C = {c1 − c2 : c1, c2 ∈ C} is syndetic provided that d(C) > 0. (To see this, pick a subset {i1, . . . , im} ⊆ Z which is maximal subject to the condition that C − i1, . . . ,C − im are mutually disjoint. This is possible since disjointness implies d(C−i1∪. . .∪C−im) = m·d(C), therefore m is at most 1/d(C). But then maximality implies that for each n ∈ Z there is some ik, k ∈ {1, . . . ,m} such that (C − n) ∩ (C − ik) , ∅, resp. n ∈ (C − C) + ik. Thus ⋃m k=1(C − C) + ik = Z.) Simple counterexamples yield that the analogous statement fails when two different sets are considered, but Renling Jin discovered the following interesting result. Theorem 1 ([Jin02]). Let A, B ⊆ Z, d(A), d(B) > 0. Then A − B is piecewise syndetic. In Section 1 we reprove Jin’s Theorem by reducing it to the C − C case. Subsequently we discuss some modifications of our argument which allow to recover the refinements resp. generalizations of Jin’s result found in [JK03, BFW06, Jin08, BBF]. 1 Ultrafilter proof of Jin’s Theorem As indicated in the abstract, we aim to show that one can shift a given large set A ⊆ Z by an ultrafilter so that it will have large intersection with another, previously specified set. We motivate the definition of this ultrafilter-shift by means of analogy: For n ∈ Z denote by e(n) the principle ultrafilter on Z which corresponds to n and notice that A − n = {k ∈ Z : n ∈ A − k} = {k ∈ Z : A − k ∈ e(n)}. Given an ultrafilter p on Z, we thus define {k ∈ Z : A − k ∈ p} as the official meaning of “A − p”. 2000 Mathematics Subject Classification. 11B05, 05D10.
منابع مشابه
Diagonalising an Ultrafilter and Preserving a P -point
With Ramsey-theoretic methods we show: It is consistent that there is a forcing that diagonalises one ultrafilter over ω and preserves another ultrafilter.
متن کاملA new proof of the 2-dimensional Halpern–Läuchli Theorem
We provide an ultrafilter proof of the 2-dimensional Halpern–Läuchli Theorem in the following sense. If T0 and T1 are trees and T0 ⊗ T1 denotes their level product, we exhibit an ultrafilter U ∈ β(T0 ⊗ T1) so that every A ∈ U contains a subset of the form S0 ⊗ S1 for suitable strong subtrees of T0 and T1. We then discuss obstacles to extending our method of proof to higher dimensions.
متن کاملThe Ultrafilter Closure in ZF
It is well known that, in a topological space, the open sets can be characterized using filter convergence. In ZF (Zermelo-Fraenkel set theory without the Axiom of Choice), we cannot replace filters by ultrafilters. It is proven that the ultrafilter convergence determines the open sets for every topological space if and only if the Ultrafilter Theorem holds. More, we can also prove that the Ult...
متن کاملIsabelle/HOL-NSA — Non-Standard Analysis
2 Filter: Filters and Ultrafilters 13 2.1 Definitions and basic properties . . . . . . . . . . . . . . . . . 13 2.1.1 Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.1.2 Ultrafilters . . . . . . . . . . . . . . . . . . . . . . . . 13 2.1.3 Free Ultrafilters . . . . . . . . . . . . . . . . . . . . . . 14 2.2 Collect properties . . . . . . . . . . . . . . . . . . . . . . . . . ...
متن کاملA Ramsey-classification Theorem and Its Application in the Tukey Theory of Ultrafilters
Motivated by a Tukey classification problem we develop here a new topological Ramsey space R1 that in its complexity comes immediately after the classical is a natural Ellentuck space [5]. Associated with R1 is an ultrafilter U1 which is weakly Ramsey but not Ramsey. We prove a canonization theorem for equivalence relations on fronts on R1. This is analogous to the Pudlak-Rödl Theorem canonizin...
متن کامل