Learning-Augmented Decentralized Online Convex Optimization in Networks
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916443457585152 |
|---|---|
| author | Li, Pengfei Yang, Jianyi Wierman, Adam Ren, Shaolei |
| author_facet | Li, Pengfei Yang, Jianyi Wierman, Adam Ren, Shaolei |
| contents | This paper studies decentralized online convex optimization in a networked multi-agent system and proposes a novel algorithm, Learning-Augmented Decentralized Online optimization (LADO), for individual agents to select actions only based on local online information. LADO leverages a baseline policy to safeguard online actions for worst-case robustness guarantees, while staying close to the machine learning (ML) policy for average performance improvement. In stark contrast with the existing learning-augmented online algorithms that focus on centralized settings, LADO achieves strong robustness guarantees in a decentralized setting. We also prove the average cost bound for LADO, revealing the tradeoff between average performance and worst-case robustness and demonstrating the advantage of training the ML policy by explicitly considering the robustness requirement. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_10158 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Learning-Augmented Decentralized Online Convex Optimization in Networks Li, Pengfei Yang, Jianyi Wierman, Adam Ren, Shaolei Machine Learning Optimization and Control This paper studies decentralized online convex optimization in a networked multi-agent system and proposes a novel algorithm, Learning-Augmented Decentralized Online optimization (LADO), for individual agents to select actions only based on local online information. LADO leverages a baseline policy to safeguard online actions for worst-case robustness guarantees, while staying close to the machine learning (ML) policy for average performance improvement. In stark contrast with the existing learning-augmented online algorithms that focus on centralized settings, LADO achieves strong robustness guarantees in a decentralized setting. We also prove the average cost bound for LADO, revealing the tradeoff between average performance and worst-case robustness and demonstrating the advantage of training the ML policy by explicitly considering the robustness requirement. |
| title | Learning-Augmented Decentralized Online Convex Optimization in Networks |
| topic | Machine Learning Optimization and Control |
| url | https://arxiv.org/abs/2306.10158 |