Publications

BibTex format

@article{Parpas:2017:10.1137/16M1082299,
author = {Parpas, P},
doi = {10.1137/16M1082299},
journal = {SIAM Journal on Scientific Computing},
pages = {S681--S701},
title = {A multilevel proximal gradient algorithm for a class of composite optimization problems},
url = {http://dx.doi.org/10.1137/16M1082299},
volume = {39},
year = {2017}
}

Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - Composite optimization models consist of the minimization of the sumof a smooth (not necessarily convex) function and a non-smooth convex function.Such models arise in many applications where, in addition to the composite natureof the objective function, a hierarchy of models is readily available. It is commonto take advantage of this hierarchy of models by first solving a low fidelity modeland then using the solution as a starting point to a high fidelity model. We adoptan optimization point of view and show how to take advantage of the availability ofa hierarchy of models in a consistent manner. We do not use the low fidelity modeljust for the computation of promising starting points but also for the computa-tion  of  search  directions.  We  establish  the  convergence  and  convergence  rate  ofthe proposed algorithm. Our numerical experiments on large scale image restora-tion problems and the transition path problem suggest that, for certain classes ofproblems, the proposed algorithm is significantly faster than the state of the art.
AU  - Parpas,P
DO  - 10.1137/16M1082299
EP  - 701
PY  - 2017///
SN  - 1095-7197
SP  - 681
TI  - A multilevel proximal gradient algorithm for a class of composite optimization problems
T2  - SIAM Journal on Scientific Computing
UR  - http://dx.doi.org/10.1137/16M1082299
UR  - http://hdl.handle.net/10044/1/48477
VL  - 39
ER  -

Download

DR PANOS PARPAS

Contact

Location

Summary

Citation

BibTex format

RIS format (EndNote, RefMan)