Search

Your search keyword '"Grigas, Paul"' showing total 3 results
Start Over Author "Grigas, Paul" Publisher springer berlin heidelberg
3 results on '"Grigas, Paul"'

2. New analysis and results for the Frank–Wolfe method

Author

Massachusetts Institute of Technology. Operations Research Center, Freund, Robert Michael, Grigas, Paul Edward, Massachusetts Institute of Technology. Operations Research Center, Freund, Robert Michael, and Grigas, Paul Edward

Abstract

We present new results for the Frank–Wolfe method (also known as the conditional gradient method). We derive computational guarantees for arbitrary step-size sequences, which are then applied to various step-size rules, including simple averaging and constant step-sizes. We also develop step-size rules and computational guarantees that depend naturally on the warm-start quality of the initial (and subsequent) iterates. Our results include computational guarantees for both duality/bound gaps and the so-called FW gaps. Lastly, we present complexity bounds in the presence of approximate computation of gradients and/or linear optimization subproblem solutions., United States. Air Force Office of Scientific Research (AFOSR Grant No. FA9550-11-1-0141), Pontifical Catholic University of Chile (MIT-Chile-Pontificia Universidad Católica de Chile Seed Fund), National Science Foundation (U.S.) (NSF Graduate Research Fellowship No. 1122374)

Published
2016

3. New analysis and results for the Frank–Wolfe method

Author

Paul Grigas, Robert M. Freund, Massachusetts Institute of Technology. Operations Research Center, Freund, Robert Michael, and Grigas, Paul Edward

Subjects
Mathematical optimization, 021103 operations research, Linear programming, General Mathematics, Computation, Numerical analysis, 0211 other engineering and technologies, Duality (optimization), 010103 numerical & computational mathematics, 02 engineering and technology, 01 natural sciences, Frank–Wolfe algorithm, Iterated function, Simple (abstract algebra), Applied mathematics, 0101 mathematics, Constant (mathematics), Software, Mathematics
Abstract

We present new results for the Frank–Wolfe method (also known as the conditional gradient method). We derive computational guarantees for arbitrary step-size sequences, which are then applied to various step-size rules, including simple averaging and constant step-sizes. We also develop step-size rules and computational guarantees that depend naturally on the warm-start quality of the initial (and subsequent) iterates. Our results include computational guarantees for both duality/bound gaps and the so-called FW gaps. Lastly, we present complexity bounds in the presence of approximate computation of gradients and/or linear optimization subproblem solutions., United States. Air Force Office of Scientific Research (AFOSR Grant No. FA9550-11-1-0141), Pontifical Catholic University of Chile (MIT-Chile-Pontificia Universidad Católica de Chile Seed Fund), National Science Foundation (U.S.) (NSF Graduate Research Fellowship No. 1122374)

Published
2013

Catalog

Books, media, physical & digital resources
See catalog results