The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
본 논문에서는 강력하게 연결된 통신 네트워크에 대한 제한된 분산 온라인 최적화 문제를 조사합니다. 여기서 각 에이전트의 로컬 비용 함수는 환경 요인으로 인해 시간에 따라 달라집니다. 우리는 불균형 방향성 네트워크를 통한 분산형 온라인 투영 하위 그라데이션 방법을 제안합니다. 제안된 방법의 성능은 시간에 따른 누적 비용과 돌이켜보면 최적 전략의 비용 사이의 오차로 정의되는 후회로 평가됩니다. 우리는 강력하게 볼록한 비용 함수에 대해 로그 후회 한계를 달성할 수 있음을 보여줍니다. 또한 확산장에 대한 분산 추정에 대한 수치적 예를 통해 제안된 방법의 타당성을 입증합니다.
Makoto YAMASHITA
Osaka University
Naoki HAYASHI
Osaka University
Takeshi HATANAKA
Tokyo Institute of Technology
Shigemasa TAKAI
Osaka University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
부
Makoto YAMASHITA, Naoki HAYASHI, Takeshi HATANAKA, Shigemasa TAKAI, "Logarithmic Regret for Distributed Online Subgradient Method over Unbalanced Directed Networks" in IEICE TRANSACTIONS on Fundamentals,
vol. E104-A, no. 8, pp. 1019-1026, August 2021, doi: 10.1587/transfun.2020EAP1111.
Abstract: This paper investigates a constrained distributed online optimization problem over strongly connected communication networks, where a local cost function of each agent varies in time due to environmental factors. We propose a distributed online projected subgradient method over unbalanced directed networks. The performance of the proposed method is evaluated by a regret which is defined by the error between the cumulative cost over time and the cost of the optimal strategy in hindsight. We show that a logarithmic regret bound can be achieved for strongly convex cost functions. We also demonstrate the validity of the proposed method through a numerical example on distributed estimation over a diffusion field.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2020EAP1111/_p
부
@ARTICLE{e104-a_8_1019,
author={Makoto YAMASHITA, Naoki HAYASHI, Takeshi HATANAKA, Shigemasa TAKAI, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Logarithmic Regret for Distributed Online Subgradient Method over Unbalanced Directed Networks},
year={2021},
volume={E104-A},
number={8},
pages={1019-1026},
abstract={This paper investigates a constrained distributed online optimization problem over strongly connected communication networks, where a local cost function of each agent varies in time due to environmental factors. We propose a distributed online projected subgradient method over unbalanced directed networks. The performance of the proposed method is evaluated by a regret which is defined by the error between the cumulative cost over time and the cost of the optimal strategy in hindsight. We show that a logarithmic regret bound can be achieved for strongly convex cost functions. We also demonstrate the validity of the proposed method through a numerical example on distributed estimation over a diffusion field.},
keywords={},
doi={10.1587/transfun.2020EAP1111},
ISSN={1745-1337},
month={August},}
부
TY - JOUR
TI - Logarithmic Regret for Distributed Online Subgradient Method over Unbalanced Directed Networks
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1019
EP - 1026
AU - Makoto YAMASHITA
AU - Naoki HAYASHI
AU - Takeshi HATANAKA
AU - Shigemasa TAKAI
PY - 2021
DO - 10.1587/transfun.2020EAP1111
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E104-A
IS - 8
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - August 2021
AB - This paper investigates a constrained distributed online optimization problem over strongly connected communication networks, where a local cost function of each agent varies in time due to environmental factors. We propose a distributed online projected subgradient method over unbalanced directed networks. The performance of the proposed method is evaluated by a regret which is defined by the error between the cumulative cost over time and the cost of the optimal strategy in hindsight. We show that a logarithmic regret bound can be achieved for strongly convex cost functions. We also demonstrate the validity of the proposed method through a numerical example on distributed estimation over a diffusion field.
ER -