Nonhomogeneous Place-dependent Markov Chains, Unsynchronised AIMD, and Optimisation-Reference-Cited by-同舟云学术

Nonhomogeneous Place-dependent Markov Chains, Unsynchronised AIMD, and Optimisation

Published:2019-08-31 Issue:4 Volume:66 Page:1-37
ISSN:0004-5411
Container-title:Journal of the ACM
language:en
Short-container-title:J. ACM

Author:

Wirth Fabian R.¹^ORCID,Stüdli Sonja²,Yu Jia Yuan³,Corless Martin⁴,Shorten Robert⁵

Affiliation:

1. University of Passau, Germany

2. University of Newcastle, Australia

3. Concordia University, Quebec, Canada

4. Purdue University, Newcastle, West Lafayette, IN, USA

5. Imperial College London and University College Dublin, Dublin, Ireland

Abstract

A stochastic algorithm is presented for a class of optimisation problems that arise when a group of agents compete to share a single constrained resource in an optimal manner. The approach uses intermittent single-bit feedback, which indicates a constraint violation and does not require inter-agent communication. The algorithm is based on a positive matrix model of AIMD, which is extended to the nonhomogeneous Markovian case. The key feature is the assignment of back-off probabilities to the individual agents as a function of the past average access to the resource. This leads to a nonhomogeneous Markov chain in an extended state space, and we show almost sure convergence of the average access to the social optimum.

Funder

Seventh Framework Programme

Science Foundation Ireland

Publisher

Association for Computing Machinery (ACM)

Subject

Artificial Intelligence,Hardware and Architecture,Information Systems,Control and Systems Engineering,Software

Link

https://dl.acm.org/doi/pdf/10.1145/3312741

Reference60 articles.

1. A Stochastic Approach to “Dynamic-Demand” Refrigerator Control

2. M. F. Barnsley S. G. Demko J. H. Elton and J. S. Geronimo. 1988. Invariant measures for Markov processes arising from iterated function systems with place-dependent probabilities. Annales de l'Institut Henri Poincaré

3. (B) Probabilités et Statistiques 24 3 (1988) 367--394. M. F. Barnsley S. G. Demko J. H. Elton and J. S. Geronimo. 1988. Invariant measures for Markov processes arising from iterated function systems with place-dependent probabilities. Annales de l'Institut Henri Poincaré

4. (B) Probabilités et Statistiques 24 3 (1988) 367--394.

5. Emulating AQM from end hosts

Cited by 9 articles. 订阅此论文施引文献订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献

1. Consensus With a Linear Constraint;IEEE Transactions on Automatic Control;2024-01

2. Communication-Efficient Allocation of Multiple Indivisible Resources in a Federated Multi-Agent System;2023 62nd IEEE Conference on Decision and Control (CDC);2023-12-13

3. On unique ergodicity of coupled AIMD flows;International Journal of Control;2023-10-02

4. Communication-efficient Preference-based Federated Multi-resource Allocation;2023 59th Annual Allerton Conference on Communication, Control, and Computing (Allerton);2023-09-26

5. Existence of a Unique Invariant Measure and Ergodic Property in AIMD-based Multi-resource Allocation;2023 American Control Conference (ACC);2023-05-31