TOPICS
Search

Mean Width


The mean width of a convex body K subset R^n is the average distance between its parallel supporting hyperplanes over uniformly distributed directions (Schneider 2014). In terms of its convex body support function h_K(u),

 w(K)=int_(S^(n-1))[h_K(u)+h_K(-u)]dsigma(u)=2int_(S^(n-1))h_K(u)dsigma(u),

where sigma is the rotation-invariant probability measure on the unit sphere. A ball of radius r has mean width 2r. For a three-dimensional convex body, this is its mean tangent diameter.

Milo (2026) obtained bounds relating mean width to the number of facets of an n-dimensional polytope P, n>=2, contained in the unit ball. Write M^*(P)=w(P)/2 and let N be the number of facets. There is a universal constant c>0 such that M^*(P)<=1-c/[ln(eN/n)]^2 when N<=e^n, and M^*(P)<=1-(c/n)N^(-2/(n-1)) for arbitrary N. The second bound is attained up to a factor of order lnn in the deficit 1-M^*(P) by suitable polytopes with at most N facets when N>=exp(Cn) for a universal constant C. Milo (2026) credits discussions with ChatGPT with developing the main proof ideas.


See also

Convex Body Support Function, Mean Tangent Diameter, Polytope

Explore with Wolfram|Alpha

References

Milo, T. "Combinatorial Dvoretzky Theory." 19 Nov 2026. https://arxiv.org/abs/2611.26253v1.Schneider, R. Convex Bodies: The Brunn-Minkowski Theory, 2nd expanded ed. Cambridge, England: Cambridge University Press, 2014. https://doi.org/10.1017/CBO9781139003858.

Cite this as:

Weisstein, Eric W. "Mean Width." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/MeanWidth.html

Subject classifications