TY - GEN
T1 - Improved Convergence Rates for Distributed Resource Allocation
AU - Nedich, Angelia
AU - Olshevsky, Alex
AU - Shi, Wei
N1 - Funding Information:
*This work has been supported by the ONR grant No. N00014-12-1-0998, the NSF grant CCF-1717391, the NSF award 1740452, and the Air Force award FA 990-15-10394.
Publisher Copyright:
© 2018 IEEE.
PY - 2018/7/2
Y1 - 2018/7/2
N2 - In this paper, we develop a class of decentralized algorithms for solving a convex resource allocation optimization problem over a connected network. By observing a connection between the resource allocation and the consensus optimization, we propose a novel class of algorithms for solving the resource allocation problem with improved convergence guarantees. Specifically, we introduce an algorithm for solving the resource allocation problem with an o(1/k) convergence rate when the agents' objective functions are generally convex and per agent local constraints are allowed; we then introduce a gradient-based algorithm for the case when per agent local constraints are absent and show that such scheme achieves geometric convergence with an improved scalability. We also provide a projection-gradient-based algorithm which can handle smooth objective and simple constraints more efficiently.
AB - In this paper, we develop a class of decentralized algorithms for solving a convex resource allocation optimization problem over a connected network. By observing a connection between the resource allocation and the consensus optimization, we propose a novel class of algorithms for solving the resource allocation problem with improved convergence guarantees. Specifically, we introduce an algorithm for solving the resource allocation problem with an o(1/k) convergence rate when the agents' objective functions are generally convex and per agent local constraints are allowed; we then introduce a gradient-based algorithm for the case when per agent local constraints are absent and show that such scheme achieves geometric convergence with an improved scalability. We also provide a projection-gradient-based algorithm which can handle smooth objective and simple constraints more efficiently.
UR - http://www.scopus.com/inward/record.url?scp=85062195407&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85062195407&partnerID=8YFLogxK
U2 - 10.1109/CDC.2018.8619322
DO - 10.1109/CDC.2018.8619322
M3 - Conference contribution
AN - SCOPUS:85062195407
T3 - Proceedings of the IEEE Conference on Decision and Control
SP - 172
EP - 177
BT - 2018 IEEE Conference on Decision and Control, CDC 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 57th IEEE Conference on Decision and Control, CDC 2018
Y2 - 17 December 2018 through 19 December 2018
ER -