4 search results - page 1 / 1 » All learning is Local: Multi-agent Learning in Global Reward... |
90
Voted
NIPS
2003
15 years 1 months ago
2003
In large multiagent games, partial observability, coordination, and credit assignment persistently plague attempts to design good learning algorithms. We provide a simple and efï¬...
96
Voted
HIS
2004
15 years 1 months ago
2004
In this paper, we describe how certain aspects of the biological phenomena of stigmergy can be imported into multiagent reinforcement learning (MARL), with the purpose of better e...
88
Voted
ICML
2003
IEEE
16 years 13 days ago
2003
IEEE
The paper explores a very simple agent design method called Q-decomposition, wherein a complex agent is built from simpler subagents. Each subagent has its own reward function and...
100
Voted
ICML
1999
IEEE
16 years 13 days ago
1999
IEEE
Many interesting problems, such as power grids, network switches, and tra c ow, that are candidates for solving with reinforcement learningRL, alsohave properties that make distri...