Uniqueness of OptISTA in composite convex minimization with a proximable term

Determine whether OptISTA is the uniquely minimax optimal deterministic first-order algorithm for composite convex minimization with a proximable term, i.e., ascertain if any other deterministic first-order method achieves the exact optimal worst-case rate on this problem class.

Background

The paper references OptISTA as achieving exact optimality for composite convex minimization with a proximable term. However, the authors explicitly state that it is unknown if OptISTA is uniquely optimal; they further note reported numerical evidence suggesting non-uniqueness.

This suggests a broader open question of characterization of all exact optimal algorithms in the composite setting, analogous to the complete characterization achieved in the paper for nonexpansive fixed-point problems.

References

To the best of our knowledge, it is unknown whether these algorithms are uniquely optimal for their respective problem classes—interestingly, for the case of OptISTA, it is reported in that numerical evidence suggests that it is not unique.

H-invariance theory: A complete characterization of minimax optimal fixed-point algorithms  (2511.14915 - Yoon et al., 18 Nov 2025) in Conclusion