Search for: All records

Abstract The three-distance theorem (also known as the three-gap theorem or Steinhaus problem) states that, for any given real number $$\alpha $$ and integer $$N$$, there are at most three values for the distances between consecutive elements of the Kronecker sequence $$\alpha , 2\alpha ,\ldots , N\alpha $$ mod 1. In this paper, we consider a natural generalization of the three-distance theorem to the higher-dimensional Kronecker sequence $$\vec \alpha , 2\vec \alpha ,\ldots , N\vec \alpha $$ modulo an integer lattice. We prove that in 2D, there are at most five values that can arise as a distance between nearest neighbors, for all choices of $$\vec \alpha $$ and $$N$$. Furthermore, for almost every $$\vec \alpha $$, five distinct distances indeed appear for infinitely many $$N$$ and hence five is the best possible general upper bound. In higher dimensions, we have similar explicit, but less precise, upper bounds. For instance, in 3D, our bound is 13, though we conjecture the truth to be 9. We furthermore study the number of possible distances from a point to its nearest neighbor in a restricted cone of directions. This may be viewed as a generalization of the gap length in 1D. For large cone angles, we use geometric arguments to produce explicit bounds directly analogous to the three-distance theorem. For small cone angles, we use ergodic theory of homogeneous flows in the space of unimodular lattices to show that the number of distinct lengths is (1) unbounded for almost all $$\vec \alpha $$ and (2) bounded for $$\vec \alpha $$ that satisfy certain Diophantine conditions. 

Recently, the first author together with Jens Marklof studied generalizations of the classical three distance theorem to higher-dimensional toral rotations, giving upper bounds in all dimensions for the corresponding numbers of distances with respect to any flat Riemannian metric. In dimension two they proved a five distance theorem, which is best possible. In this paper, we establish analogous bounds, in all dimensions, for the maximum metric. We also show that in dimensions two and three our bounds are best possible.