Linear Size Distance Preservers
نویسنده
چکیده
The famous shortest path tree lemma states that, for any node s in a graph G = (V,E), there is a subgraph on O(n) edges that preserves all distances between node pairs in the set {s}×V . A very basic question in distance sketching research, with applications to other problems in the field, is to categorize when else graphs admit sparse subgraphs that preserve distances between a set P of p node pairs, where P has some different structure than {s}×V or possibly no guaranteed structure at all. Trivial lower bounds of a path or a clique show that such a subgraph will need Ω(n + p) edges in the worst case. The question is then to determine when these trivial lower bounds are sharp; that is, when do graphs have linear size distance preservers on O(n+ p) edges? In this paper, we make the first new progress on this fundamental question in over ten years. We show: 1. All G,P has a distance preserver onO(n) edges whenever p = O(n), even if G is directed and/or weighted. These are the first nontrivial preservers of size O(n) known for directed graphs. 2. All G,P has a distance preserver on O(p) edges whenever p = Ω (
منابع مشابه
Euclidean and circum-Euclidean distance matrices: Characterizations and linear preservers
Short proofs are given to various characterizations of the (circum-)Euclidean squared distance matrices. Linear preserver problems related to these matrices are discussed.
متن کاملEla Euclidean and Circum-euclidean Distance Matrices: Characterizations and Linear Preservers
Short proofs are given to various characterizations of the (circum-)Euclidean squared distance matrices. Linear preserver problems related to these matrices are discussed.
متن کاملOn multiplicative (strong) linear preservers of majorizations
In this paper, we study some kinds of majorizations on $textbf{M}_{n}$ and their linear or strong linear preservers. Also, we find the structure of linear or strong linear preservers which are multiplicative, i.e. linear or strong linear preservers like $Phi $ with the property $Phi (AB)=Phi (A)Phi (B)$ for every $A,Bin textbf{M}_{n}$.
متن کاملLinear Preservers of Majorization
For vectors $X, Yin mathbb{R}^{n}$, we say $X$ is left matrix majorized by $Y$ and write $X prec_{ell} Y$ if for some row stochastic matrix $R, ~X=RY.$ Also, we write $Xsim_{ell}Y,$ when $Xprec_{ell}Yprec_{ell}X.$ A linear operator $Tcolon mathbb{R}^{p}to mathbb{R}^{n}$ is said to be a linear preserver of a given relation $prec$ if $Xprec Y$ on $mathbb{R}^{p}$ implies that $TXprec TY$ on $mathb...
متن کاملSGLT-MAJORIZATION ON Mn,m AND ITS LINEAR PRESERVERS
A matrix R is said to be g-row substochastic if Re ≤ e. For X, Y ∈ Mn,m, it is said that X is sglt-majorized by Y , X ≺sglt Y , if there exists an n-by-n lower triangular g-row substochastic matrix R such that X = RY . This paper characterizes all (strong) linear preservers and strong linear preservers of ≺sglt on Rn and Mn,m, respectively.
متن کامل