and another one is a proper, convex and lower semicontinuous function. We present the \( O\left(1/k^2 \right) \) convergence rate under three classical rules proposed by Nesterov, Chambolle-Dossal and Attouch-Cabot without the strong convexity, which is compatible with the results of the continuous-time dynamic system.
We further present a primal-dual algorithm based on the explicit discretization when both objective functions are continuously differentiable convex functions.
We show the corresponding non-ergodic convergence rate and prove that the sequence of iterates generated by the primal-dual algorithm weakly converges to a primal-dual optimal solution.