TOPICS
Search

Mean Dimension


Mean dimension of a compact metrizable dynamical system measures the covering dimension required per orbit coordinate. For a homeomorphism h:X->X, it is

 mdim(X,h)=sup_(U)lim_(N->infty)(D( v _(j=0)^(N-1)h^(-j)U))/N,

where U ranges over finite open covers and D(V) is the least order of an open refinement. The order is one less than the largest number of members with a nonempty common intersection, and  v is the common refinement by intersections. The limit exists by subadditivity. Finite-dimensional compact systems have mean dimension zero, while the supremum may be infinite for an infinite-dimensional space.

Minimality means that every orbit is dense.


See also

Lebesgue Covering Dimension

Explore with Wolfram|Alpha

References

Lindenstrauss, E. "Mean Dimension, Small Entropy Factors and an Embedding Theorem." Publ. Math. IHES 89, 227-262, 1999. https://doi.org/10.1007/BF02698858.

Cite this as:

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

Subject classifications