ε-net


Definition Suppose X is a metric space with a metric d, and suppose S is a subset of X. Let ε be a positive real number. A subset NS is an ε-net for S if, for all xS, there is an yN, such that d(x,y)<ε.

For any ε>0 and SX, the set S is trivially an ε-net for itself.

TheoremMathworldPlanetmath Suppose X is a metric space with a metric d, and suppose S is a subset of X. Let ε be a positive real number. Then N is an ε-net for S, if and only if

{Bε(y)yN}

is a cover for S. (Here Bε(x) is the open ball with center x and radius ε.)

Proof. Suppose N is an ε-net for S. If xS, there is an yN such that xBε(y). Thus, x is covered by some set in {Bε(x)xN}. Conversely, suppose {Bε(y)yN} is a cover for S, and suppose xS. By assumptionPlanetmathPlanetmath, there is an yN, such that xBε(y). Hence d(x,y)<ε with yN.

Example In X=2 with the usual Cartesian metric, the set

N={(a,b)a,b }

is an ε-net for X assuming that ε>2/2.

The above definition and example can be found in [1], page 64-65.

References

Title ε-net
Canonical name varepsilonnet
Date of creation 2013-03-22 13:37:54
Last modified on 2013-03-22 13:37:54
Owner Koro (127)
Last modified by Koro (127)
Numerical id 4
Author Koro (127)
Entry type Definition
Classification msc 54E35
Related topic Cover