Embedding multidimensional grids into optimal hypercubes
Abstract
Let G and H be graphs, with |V(H)|≥ |V(G)| , and f:V(G)→ V(H) a one to one map of their vertices. Let dilation(f) = max\ distH(f(x),f(y)): xy∈ E(G) \, where distH(v,w) is the distance between vertices v and w of H. Now let B(G,H) = minf\ dilation(f) \, over all such maps f. The parameter B(G,H) is a generalization of the classic and well studied "bandwidth" of G, defined as B(G,P(n)), where P(n) is the path on n points and n = |V(G)|. Let [a1× a2× ·s × ak ] be the k-dimensional grid graph with integer values 1 through ai in the i'th coordinate. In this paper, we study B(G,H) in the case when G = [a1× a2× ·s × ak ] and H is the hypercube Qn of dimension n = log2(|V(G)|) , the hypercube of smallest dimension having at least as many points as G. Our main result is that B( [a1× a2× ·s × ak ],Qn) 3k, provided ai ≥ 222 for each 1 i k. For such G, the bound 3k improves on the previous best upper bound 4k+O(1). Our methods include an application of Knuth's result on two-way rounding and of the existence of spanning regular cyclic caterpillars in the hypercube.
Turn this paper into a full lesson
ArcXiv compiles a staged curriculum from this paper: 8-12 lessons across beginner → advanced, synthesised section guides, visuals, flashcards, a quiz, exercises, and on-demand deep dives per section. Grounded in the abstract, never invented.