We study the worst-case complexity of a non-monotone line search framework that covers a wide variety of known techniques published in the literature. In this framework, the non-monotonicity is controlled by a sequence of nonnegative parameters. We obtain complexity bounds to achieve approximate first-order optimality even when this sequence is not summable.
Cite
@article{arxiv.1911.06072,
title = {A Generalized Worst-Case Complexity Analysis for Non-Monotone Line Searches},
author = {Geovani N. Grapiglia and Ekkehard W. Sachs},
journal= {arXiv preprint arXiv:1911.06072},
year = {2021}
}