Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
5773851 | Journal of Complexity | 2017 | 16 Pages |
Abstract
We obtain a new lower bound on the information-based complexity of first-order minimization of smooth and convex functions. We show that the bound matches the worst-case performance of the recently introduced Optimized Gradient Method (Drori and Teboulle, 2013; Kim and Fessler, 2015), thereby establishing that the bound is tight and can be realized by an efficient algorithm. The proof is based on a novel construction technique of smooth and convex functions.
Related Topics
Physical Sciences and Engineering
Mathematics
Analysis
Authors
Yoel Drori,