Non-Stationary Policy Learning for Multi-Timescale Multi-Agent Reinforcement Learning
Emami, Patrick, Zhang, Xiangyu, Biagioni, David, Zamzam, Ahmed S.
–arXiv.org Artificial Intelligence
In multi-timescale multi-agent reinforcement learning (MARL), agents interact across different timescales. In general, policies for time-dependent behaviors, such as those induced by multiple timescales, are non-stationary. Learning non-stationary policies is challenging and typically requires sophisticated or inefficient algorithms. Motivated by the prevalence of this control problem in real-world complex systems, we introduce a simple framework for learning non-stationary policies for multi-timescale MARL. Our approach uses available information about agent timescales to define a periodic time encoding. In detail, we theoretically demonstrate that the effects of non-stationarity introduced by multiple timescales can be learned by a periodic multi-agent policy. To learn such policies, we propose a policy gradient algorithm that parameterizes the actor and critic with phase-functioned neural networks, which provide an inductive bias for periodicity. The framework's ability to effectively learn multi-timescale policies is validated on a gridworld and building energy management environment.
arXiv.org Artificial Intelligence
Jul-17-2023
- Country:
- North America > United States
- California > Los Angeles County
- Long Beach (0.04)
- Colorado
- Broomfield County > Broomfield (0.04)
- Jefferson County > Golden (0.04)
- Louisiana > Orleans Parish
- New Orleans (0.04)
- California > Los Angeles County
- North America > United States
- Genre:
- Research Report (0.64)
- Industry:
- Energy
- Power Industry (0.66)
- Renewable (0.68)
- Government > Regional Government
- Energy