An official website of the United States government
We propose a novel policy gradient method for multi-agent reinforcement learning, which leverages two different variance-reduction techniques and does not require large batches over iterations. Specifically, we propose a momentum-based decentralized policy gradient tracking (MDPGT) where a new momentum-based variance reduction technique is used to approximate the local policy gradient surrogate with importance sampling, and an intermediate parameter is adopted to track two consecutive policy gradient surrogates. MDPGT provably achieves the best available sample complexity of O(N -1 e -3) for converging to an e-stationary point of the global average of N local performance functions (possibly nonconcave). This outperforms the state-of-the-art sample complexity in decentralized model-free reinforcement learning and when initialized with a single trajectory, the sample complexity matches those obtained by the existing decentralized policy gradient methods. We further validate the theoretical claim for the Gaussian policy function. When the required error tolerance e is small enough, MDPGT leads to a linear speed up, which has been previously established in decentralized stochastic optimization, but not for reinforcement learning. Lastly, we provide empirical results on a multi-agent reinforcement learning benchmark environment to support our theoretical findings.
more »
« less
Fast multi-agent temporal-difference learning via homotopy stochastic primal-dual method
We study a distributed policy evaluation problem in which a group of agents with jointly observed states and private local actions and rewards collaborate to learn the value function of a given policy via local computation and communication. This problem arises in various large-scale multi-agent systems, including power grids, intelligent transportation systems, wireless sensor networks, and multi-agent robotics. We develop and analyze a new distributed temporal-difference learning algorithm that minimizes the mean-square projected Bellman error. Our approach is based on a stochastic primal-dual method and we improve the best-known convergence rate from $$O(1/\sqrt{T})$$ to $O(1/T)$, where $$T$$ is the total number of iterations. Our analysis explicitly takes into account the Markovian nature of the sampling and addresses a broader class of problems than the commonly-used i.i.d. sampling scenario.
more »
« less
- PAR ID:
- 10128755
- Date Published:
- Journal Name:
- Optimization Foundations for Reinforcement Learning Workshop, 33rd Conference on Neural Information Processing Systems
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
We study the multi-agent multi-armed bandit (MAMAB) problem, where agents are factored into overlapping groups. Each group represents a hyperedge, forming a hypergraph over the agents. At each round of interaction, the learner pulls a joint arm (composed of individual arms for each agent) and receives a reward according to the hypergraph structure. Specifically, we assume there is a local reward for each hyperedge, and the reward of the joint arm is the sum of these local rewards. Previous work introduced the multi-agent Thompson sampling (MATS) algorithm and derived a Bayesian regret bound. However, it remains an open problem how to derive a frequentist regret bound for Thompson sampling in this multi-agent setting. To address these issues, we propose an efficient variant of MATS, the epsilon-exploring Multi-Agent Thompson Sampling (eps-MATS) algorithm, which performs MATS exploration with probability epsilon while adopts a greedy policy otherwise. We prove that eps-MATS achieves a worst-case frequentist regret bound that is sublinear in both the time horizon and the local arm size. We also derive a lower bound for this setting, which implies our frequentist regret upper bound is optimal up to constant and logarithm terms, when the hypergraph is sufficiently sparse. Thorough experiments on standard MAMAB problems demonstrate the superior performance and the improved computational efficiency of eps-MATS compared with existing algorithms in the same setting.more » « less
-
We study the multi-agent multi-armed bandit (MAMAB) problem, where agents are factored into overlapping groups. Each group represents a hyperedge, forming a hypergraph over the agents. At each round of interaction, the learner pulls a joint arm (composed of individual arms for each agent) and receives a reward according to the hypergraph structure. Specifically, we assume there is a local reward for each hyperedge, and the reward of the joint arm is the sum of these local rewards. Previous work introduced the multi-agent Thompson sampling (MATS) algorithm and derived a Bayesian regret bound. However, it remains an open problem how to derive a frequentist regret bound for Thompson sampling in this multi-agent setting. To address these issues, we propose an efficient variant of MATS, the epsilon-exploring Multi-Agent Thompson Sampling (eps-MATS) algorithm, which performs MATS exploration with probability epsilon while adopts a greedy policy otherwise. We prove that eps-MATS achieves a worst-case frequentist regret bound that is sublinear in both the time horizon and the local arm size. We also derive a lower bound for this setting, which implies our frequentist regret upper bound is optimal up to constant and logarithm terms, when the hypergraph is sufficiently sparse. Thorough experiments on standard MAMAB problems demonstrate the superior performance and the improved computational efficiency of eps-MATS compared with existing algorithms in the same setting.more » « less
-
This paper proves an impossibility result for stochastic network utility maximization for multi-user wireless systems, including multi-access and broadcast systems. Every time slot an access point observes the current channel states and opportunistically selects a vector of transmission rates. Channel state vectors are assumed to be independent and identically distributed with an unknown probability distribution. The goal is to learn to make decisions over time that maximize a concave utility function of the running time average transmission rate of each user. Recently it was shown that a stochastic Frank-Wolfe algorithm converges to utility-optimality with an error of O(log(T)/T ), where T is the time the algorithm has been running. An existing O(1/T ) converse is known. The current paper improves the converse to O(log(T)/T), which matches the known achievability result. The proof uses a reduction from the opportunistic scheduling problem to a Bernoulli estimation problem. Along the way, it refines a result on Bernoulli estimation.more » « less
-
We present a closed-loop multi-arm motion planner that is scalable and flexible with team size. Traditional multi-arm robotic systems have relied on centralized motion planners, whose run times often scale exponentially with team size, and thus, fail to handle dynamic environments with open-loop control. In this paper, we tackle this problem with multi-agent reinforcement learning, where a shared policy network is trained to control each individual robot arm to reach its target end-effector pose given observations of its workspace state and target end-effector pose. The policy is trained using Soft Actor-Critic with expert demonstrations from a sampling-based motion planning algorithm (i.e., BiRRT). By leveraging classical planning algorithms, we can improve the learning efficiency of the reinforcement learning algorithm while retaining the fast inference time of neural networks. The resulting policy scales sub-linearly and can be deployed on multi-arm systems with variable team sizes. Thanks to the closed-loop and decentralized formulation, our approach generalizes to 5-10 multiarm systems and dynamic moving targets (>90% success rate for a 10-arm system), despite being trained on only 1-4 arm planning tasks with static targets.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
