Speaker
Yuri Malykhin
(Steklov Mathematical Institute of RAS, Moscow)
Description
The classical notion of Kolmogorov width of a set in a normed space measures the error of approximation of this set by n-dimensional linear subspaces. Here we consider the ''worst-case'' error of approximation.
If we take the ''average-case'' error instead, we arrive to the notion of average Kolmogorov width. We will discuss some new bounds for the average widths and some applications for the classical Kolmogorov widths.
Primary author
Yuri Malykhin
(Steklov Mathematical Institute of RAS, Moscow)
