BriefGPT.xyz
Nov, 2017
通信对非合作式多玩家多臂赌博问题的影响
The Effect of Communication on Noncooperative Multiplayer Multi-Armed Bandit Problems
HTML
PDF
Noyan Evirgen, Alper Kose
TL;DR
本研究考虑了多个玩家之间，通过Erdos-Renyi图，以不同的通信概率下的去中心化随机多臂赌博问题，使用UCB1、epsilon-Greedy和Thompson Sampling算法探究了玩家之间的连接度对累计遗憾的影响。
Abstract
We consider
decentralized stochastic
multi-armed bandit
problem with multiple players in the case of different
communication probabilities
→