Imperial College London

DrPhilippThomas

Faculty of Natural SciencesDepartment of Mathematics

Senior Lecturer in Biomathematics
 
 
 
//

Contact

 

+44 (0)20 7594 2647p.thomas

 
 
//

Location

 

626Huxley BuildingSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Kuntz:2021:10.1137/19M1268847,
author = {Kuntz, Nussio J and Thomas, P and Stan, G and Barahona, M},
doi = {10.1137/19M1268847},
journal = {SIAM Journal on Optimization},
pages = {604--625},
title = {Approximations of countably-infinite linear programs over bounded measure spaces},
url = {http://dx.doi.org/10.1137/19M1268847},
volume = {31},
year = {2021}
}

RIS format (EndNote, RefMan)

TY  - JOUR
AB - We study a class of countably-infinite-dimensional linear programs (CILPs)whose feasible sets are bounded subsets of appropriately defined spaces ofmeasures. The optimal value, optimal points, and minimal points of these CILPscan be approximated by solving finite-dimensional linear programs. We show howto construct finite-dimensional programs that lead to approximations witheasy-to-evaluate error bounds, and we prove that the errors converge to zero asthe size of the finite-dimensional programs approaches that of the originalproblem. We discuss the use of our methods in the computation of the stationarydistributions, occupation measures, and exit distributions of Markov~chains.
AU - Kuntz,Nussio J
AU - Thomas,P
AU - Stan,G
AU - Barahona,M
DO - 10.1137/19M1268847
EP - 625
PY - 2021///
SN - 1052-6234
SP - 604
TI - Approximations of countably-infinite linear programs over bounded measure spaces
T2 - SIAM Journal on Optimization
UR - http://dx.doi.org/10.1137/19M1268847
UR - http://hdl.handle.net/10044/1/85949
VL - 31
ER -