Stepsize Choice for Korpelevich’s and Popov’s Extragradient Algorithms for Convex-Concave Minimax Problems.

We show that the choice of stepsize in Korpelevich’s extragradient algorithm is sharp, while the choice of stepsize in Popov’s extragradient algorithm can be relaxed. We also extend Korpelevich’s extragradient algorithm and Popov’s extragradient algorithm (with larger stepsize) to the infinite-dimensional Hilbert space framework, with weak convergence. [ABSTRACT FROM AUTHOR]