Q-learning is widely employed for optimizing various large-dimensional networks with unknown system dynamics. Recent advancements include multi-environment mixed Q-learning (MEMQ) algorithms, which utilize multiple independent Q-learning algorithms across multiple, structurally related but distinct environments and outperform several state-of-the-art Q-learning algorithms in terms of accuracy, complexity, and robustness. We herein conduct a comprehensive probabilistic coverage analysis to ensure optimal data coverage conditions for MEMQ algorithms. First, we derive upper and lower bounds on the expectation and variance of different coverage coefficients (CC) for MEMQ algorithms. Leveraging these bounds, we develop a simple way of comparing the utilities of multiple environments in MEMQ algorithms. This approach appears to be near optimal versus our previously proposed partial ordering approach. We also present a novel CC-based MEMQ algorithm to improve the accuracy and complexity of existing MEMQ algorithms. Numerical experiments are conducted using random network graphs with four different graph properties. Our algorithm can reduce the average policy error (APE) by 65% compared to partial ordering and is 95% faster than the exhaustive search. It also achieves 60% less APE than several state-of-the-art reinforcement learning and prior MEMQ algorithms. Additionally, we numerically verify the theoretical results and show their scalability with the action-space size.
Reinforcement Learning (RL) algorithms are known to suffer from the curse of dimensionality, which refers to the fact that large-scale problems often lead to exponentially high sample complexity. A common solution is to use deep neural networks for function approximation; however, such approaches typically lack theoretical guarantees. To provably address the curse of dimensionality, we observe that many real-world problems exhibit task-specific model structures that, when properly leveraged, can improve the sample efficiency of RL. Building on this insight, we propose overcoming the curse of dimensionality by approximately factorizing the original Markov decision processes (MDPs) into smaller, independently evolving MDPs. This factorization enables the development of sample-efficient RL algorithms in both model-based and model-free settings, with the latter involving a variant of variance-reduced Q-learning. We provide improved sample complexity guarantees for both proposed algorithms. Notably, by leveraging model structure through the approximate factorization of the MDP, the dependence of sample complexity on the size of the state-action space can be exponentially reduced. Numerically, we demonstrate the practicality of our proposed methods through experiments on both synthetic MDP tasks and a wind farm-equipped storage control problem.
Offline Reinforcement Learning (RL) has emerged as a powerful alternative to imitation learning for behavior modeling in various domains, particularly in complex navigation tasks. An existing challenge with Offline RL is the signal-to-noise ratio, i.e. how to mitigate incorrect policy updates due to errors in value estimates. Towards this, multiple works have demonstrated the advantage of hierarchical offline RL methods, which decouples high-level path planning from low-level path following. In this work, we present a novel hierarchical transformer-based approach leveraging a learned quantizer of the space. This quantization enables the training of a simpler zone-conditioned low-level policy and simplifies planning, which is reduced to discrete autoregressive prediction. Among other benefits, zone-level reasoning in planning enables explicit trajectory stitching rather than implicit stitching based on noisy value function estimates. By combining this transformer-based planner with recent advancements in offline RL, our proposed approach achieves state-of-the-art results in complex long-distance navigation environments.
The adaptation to users' preferences and the ability to infer and interpret humans' beliefs and intents, which is known as the Theory of Mind (ToM), are two crucial aspects for achieving effective human-robot collaboration. Despite its importance, very few studies have investigated the impact of adaptive robots with ToM abilities. In this work, we present an exploratory comparative study to investigate how social robots equipped with ToM abilities impact users' performance and perception. We design a two-layer architecture. The Q-learning agent on the first layer learns the robot's higher-level behaviour. On the second layer, a heuristic-based ToM infers the user's intended strategy and is responsible for implementing the robot's assistance, as well as providing the motivation behind its choice. We conducted a user study in a real-world setting, involving 56 participants who interacted with either an adaptive robot capable of ToM, or with a robot lacking such abilities. Our findings suggest that participants in the ToM condition performed better, accepted the robot's assistance more often, and perceived its ability to adapt, predict and recognise their intents to a higher degree. Our preliminary insights could inform future research and pave the way for designing more complex computation architectures for adaptive behaviour with ToM capabilities.
This study presents a novel computer system performance optimization and adaptive workload management scheduling algorithm based on Q-learning. In modern computing environments, characterized by increasing data volumes, task complexity, and dynamic workloads, traditional static scheduling methods such as Round-Robin and Priority Scheduling fail to meet the demands of efficient resource allocation and real-time adaptability. By contrast, Q-learning, a reinforcement learning algorithm, continuously learns from system state changes, enabling dynamic scheduling and resource optimization. Through extensive experiments, the superiority of the proposed approach is demonstrated in both task completion time and resource utilization, outperforming traditional and dynamic resource allocation (DRA) algorithms. These findings are critical as they highlight the potential of intelligent scheduling algorithms based on reinforcement learning to address the growing complexity and unpredictability of computing environments. This research provides a foundation for the integration of AI-driven adaptive scheduling in future large-scale systems, offering a scalable, intelligent solution to enhance system performance, reduce operating costs, and support sustainable energy consumption. The broad applicability of this approach makes it a promising candidate for next-generation computing frameworks, such as edge computing, cloud computing, and the Internet of Things.
Offline reinforcement learning can enable policy learning from pre-collected, sub-optimal datasets without online interactions. This makes it ideal for real-world robots and safety-critical scenarios, where collecting online data or expert demonstrations is slow, costly, and risky. However, most existing offline RL works assume the dataset is already labeled with the task rewards, a process that often requires significant human effort, especially when ground-truth states are hard to ascertain (e.g., in the real-world). In this paper, we build on prior work, specifically RL-VLM-F, and propose a novel system that automatically generates reward labels for offline datasets using preference feedback from a vision-language model and a text description of the task. Our method then learns a policy using offline RL with the reward-labeled dataset. We demonstrate the system's applicability to a complex real-world robot-assisted dressing task, where we first learn a reward function using a vision-language model on a sub-optimal offline dataset, and then we use the learned reward to employ Implicit Q learning to develop an effective dressing policy. Our method also performs well in simulation tasks involving the manipulation of rigid and deformable objects, and significantly outperform baselines such as behavior cloning and inverse RL. In summary, we propose a new system that enables automatic reward labeling and policy learning from unlabeled, sub-optimal offline datasets.
Deep reinforcement learning (DRL) has been extensively applied to Multi-Unmanned Aerial Vehicle (UAV) network (MUN) to effectively enable real-time adaptation to complex, time-varying environments. Nevertheless, most of the existing works assume a stationary user distribution (UD) or a dynamic one with predicted patterns. Such considerations may make the UD-specific strategies insufficient when a MUN is deployed in unknown environments. To this end, this paper investigates distributed user connectivity maximization problem in a MUN with generalization to arbitrary UDs. Specifically, the problem is first formulated into a time-coupled combinatorial nonlinear non-convex optimization with arbitrary underlying UDs. To make the optimization tractable, a multi-agent CNN-enhanced deep Q learning (MA-CDQL) algorithm is proposed. The algorithm integrates a ResNet-based CNN to the policy network to analyze the input UD in real time and obtain optimal decisions based on the extracted high-level UD features. To improve the learning efficiency and avoid local optimums, a heatmap algorithm is developed to transform the raw UD to a continuous density map. The map will be part of the true input to the policy network. Simulations are conducted to demonstrate the efficacy of UD heatmaps and the proposed algorithm in maximizing user connectivity as compared to K-means methods.
Value-based reinforcement learning (RL) can in principle learn effective policies for a wide range of multi-turn problems, from games to dialogue to robotic control, including via offline RL from static previously collected datasets. However, despite the widespread use of policy gradient methods to train large language models for single turn tasks (e.g., question answering), value-based methods for multi-turn RL in an off-policy or offline setting have proven particularly challenging to scale to the setting of large language models. This setting requires effectively leveraging pretraining, scaling to large architectures with billions of parameters, and training on large datasets, all of which represent major challenges for current value-based RL methods. In this work, we propose a novel offline RL algorithm that addresses these drawbacks, casting Q-learning as a modified supervised fine-tuning (SFT) problem where the probabilities of tokens directly translate to Q-values. In this way we obtain an algorithm that smoothly transitions from maximizing the likelihood of the data during pretraining to learning a near-optimal Q-function during finetuning. Our algorithm has strong theoretical foundations, enjoying performance bounds similar to state-of-the-art Q-learning methods, while in practice utilizing an objective that closely resembles SFT. Because of this, our approach can enjoy the full benefits of the pretraining of language models, without the need to reinitialize any weights before RL finetuning, and without the need to initialize new heads for predicting values or advantages. Empirically, we evaluate our method on both pretrained LLMs and VLMs, on a variety of tasks including both natural language dialogue and robotic manipulation and navigation from images.
We provide abstract, general and highly uniform rates of asymptotic regularity for a generalized stochastic Halpern-style iteration, which incorporates a second mapping in the style of a Krasnoselskii-Mann iteration. This iteration is general in two ways: First, it incorporates stochasticity in a completely abstract way rather than fixing a sampling method; secondly, it includes as special cases stochastic versions of various schemes from the optimization literature, including Halpern's iteration as well as a Krasnoselskii-Mann iteration with Tikhonov regularization terms in the sense of Bo\c{t}, Csetnek and Meier. For these particular cases, we in particular obtain linear rates of asymptotic regularity, matching (or improving) the currently best known rates for these iterations in stochastic optimization, and quadratic rates of asymptotic regularity are obtained in the context of inner product spaces for the general iteration. We utilize these rates to give bounds on the oracle complexity of such iterations under suitable variance assumptions and batching strategies, again presented in an abstract style. Finally, we sketch how the schemes presented here can be instantiated in the context of reinforcement learning to yield novel methods for Q-learning.
Safe reinforcement learning (RL) is crucial for real-world applications, and multi-agent interactions introduce additional safety challenges. While Probabilistic Logic Shields (PLS) has been a powerful proposal to enforce safety in single-agent RL, their generalizability to multi-agent settings remains unexplored. In this paper, we address this gap by conducting extensive analyses of PLS within decentralized, multi-agent environments, and in doing so, propose Shielded Multi-Agent Reinforcement Learning (SMARL) as a general framework for steering MARL towards norm-compliant outcomes. Our key contributions are: (1) a novel Probabilistic Logic Temporal Difference (PLTD) update for shielded, independent Q-learning, which incorporates probabilistic constraints directly into the value update process; (2) a probabilistic logic policy gradient method for shielded PPO with formal safety guarantees for MARL; and (3) comprehensive evaluation across symmetric and asymmetrically shielded $n$-player game-theoretic benchmarks, demonstrating fewer constraint violations and significantly better cooperation under normative constraints. These results position SMARL as an effective mechanism for equilibrium selection, paving the way toward safer, socially aligned multi-agent systems.
A computational problem in biological reward-based learning is how credit assignment is performed in the nucleus accumbens (NAc). Much research suggests that NAc dopamine encodes temporal-difference (TD) errors for learning value predictions. However, dopamine is synchronously distributed in regionally homogeneous concentrations, which does not support explicit credit assignment (like used by backpropagation). It is unclear whether distributed errors alone are sufficient for synapses to make coordinated updates to learn complex, nonlinear reward-based learning tasks. We design a new deep Q-learning algorithm, Artificial Dopamine, to computationally demonstrate that synchronously distributed, per-layer TD errors may be sufficient to learn surprisingly complex RL tasks. We empirically evaluate our algorithm on MinAtar, the DeepMind Control Suite, and classic control tasks, and show it often achieves comparable performance to deep RL algorithms that use backpropagation.
We study the problem of estimating the \emph{value} of the largest mean among K distributions via samples from them (rather than estimating \emph{which} distribution has the largest mean), which arises from various machine learning tasks including Q-learning and Monte Carlo Tree Search (MCTS). While there have been a few proposed algorithms, their performance analyses have been limited to their biases rather than a precise error metric. In this paper, we propose a novel algorithm called HAVER (Head AVERaging) and analyze its mean squared error. Our analysis reveals that HAVER has a compelling performance in two respects. First, HAVER estimates the maximum mean as well as the oracle who knows the identity of the best distribution and reports its sample mean. Second, perhaps surprisingly, HAVER exhibits even better rates than this oracle when there are many distributions near the best one. Both of these improvements are the first of their kind in the literature, and we also prove that the naive algorithm that reports the largest empirical mean does not achieve these bounds. Finally, we confirm our theoretical findings via numerical experiments where we implement HAVER in bandit, Q-learning, and MCTS algorithms. In these experiments, HAVER consistently outperforms the baseline methods, demonstrating its effectiveness across different applications.
We study the convergence of $q$-learning and related algorithms introduced by Jia and Zhou (J. Mach. Learn. Res., 24 (2023), 161) for controlled diffusion processes. Under suitable conditions on the growth and regularity of the model parameters, we provide a quantitative error and regret analysis of both the exploratory policy improvement algorithm and the $q$-learning algorithm.
Nanorobots are a promising development in targeted drug delivery and the treatment of neurological disorders, with potential for crossing the blood-brain barrier (BBB). These small devices leverage advancements in nanotechnology and bioengineering for precise navigation and targeted payload delivery, particularly for conditions like brain tumors, Alzheimer's disease, and Parkinson's disease. Recent progress in artificial intelligence (AI) and machine learning (ML) has improved the navigation and effectiveness of nanorobots, allowing them to detect and interact with cancer cells through biomarker analysis. This study presents a new reinforcement learning (RL) framework for optimizing nanorobot navigation in complex biological environments, focusing on cancer cell detection by analyzing the concentration gradients of surrounding biomarkers. We utilize a computer simulation model to explore the behavior of nanorobots in a three-dimensional space with cancer cells and biological barriers. The proposed method uses Q-learning to refine movement strategies based on real-time biomarker concentration data, enabling nanorobots to autonomously navigate to cancerous tissues for targeted drug delivery. This research lays the groundwork for future laboratory experiments and clinical applications, with implications for personalized medicine and less invasive cancer treatments. The integration of intelligent nanorobots could revolutionize therapeutic strategies, reducing side effects and enhancing treatment effectiveness for cancer patients. Further research will investigate the practical deployment of these technologies in medical settings, aiming to unlock the full potential of nanorobotics in healthcare.
In Markov decision processes (MDPs), quantile risk measures such as Value-at-Risk are a standard metric for modeling RL agents' preferences for certain outcomes. This paper proposes a new Q-learning algorithm for quantile optimization in MDPs with strong convergence and performance guarantees. The algorithm leverages a new, simple dynamic program (DP) decomposition for quantile MDPs. Compared with prior work, our DP decomposition requires neither known transition probabilities nor solving complex saddle point equations and serves as a suitable foundation for other model-free RL algorithms. Our numerical results in tabular domains show that our Q-learning algorithm converges to its DP variant and outperforms earlier algorithms.
Path planning in high-dimensional spaces poses significant challenges, particularly in achieving both time efficiency and a fair success rate. To address these issues, we introduce a novel path-planning algorithm, Zonal RL-RRT, that leverages kd-tree partitioning to segment the map into zones while addressing zone connectivity, ensuring seamless transitions between zones. By breaking down the complex environment into multiple zones and using Q-learning as the high-level decision-maker, our algorithm achieves a 3x improvement in time efficiency compared to basic sampling methods such as RRT and RRT* in forest-like maps. Our approach outperforms heuristic-guided methods like BIT* and Informed RRT* by 1.5x in terms of runtime while maintaining robust and reliable success rates across 2D to 6D environments. Compared to learning-based methods like NeuralRRT* and MPNetSMP, as well as the heuristic RRT*J, our algorithm demonstrates, on average, 1.5x better performance in the same environments. We also evaluate the effectiveness of our approach through simulations of the UR10e arm manipulator in the MuJoCo environment. A key observation of our approach lies in its use of zone partitioning and Reinforcement Learning (RL) for adaptive high-level planning allowing the algorithm to accommodate flexible policies across diverse environments, making it a versatile tool for advanced path planning.
Link adaptation (LA) is an essential function in modern wireless communication systems that dynamically adjusts the transmission rate of a communication link to match time- and frequency-varying radio link conditions. However, factors such as user mobility, fast fading, imperfect channel quality information, and aging of measurements make the modeling of LA challenging. To bypass the need for explicit modeling, recent research has introduced online reinforcement learning (RL) approaches as an alternative to the more commonly used rule-based algorithms. Yet, RL-based approaches face deployment challenges, as training in live networks can potentially degrade real-time performance. To address this challenge, this paper considers offline RL as a candidate to learn LA policies with minimal effects on the network operation. We propose three LA designs based on batch-constrained deep Q-learning, conservative Q-learning, and decision transformer. Our results show that offline RL algorithms can match the performance of state-of-the-art online RL methods when data is collected with a proper behavioral policy.
Motivated by engineering applications such as resource allocation in networks and inventory systems, we consider average-reward Reinforcement Learning with unbounded state space and reward function. Recent works studied this problem in the actor-critic framework and established finite sample bounds assuming access to a critic with certain error guarantees. We complement their work by studying Temporal Difference (TD) learning with linear function approximation and establishing finite-time bounds with the optimal $\mathcal{O}\left(1/\epsilon^2\right)$ sample complexity. These results are obtained using the following general-purpose theorem for non-linear Stochastic Approximation (SA). Suppose that one constructs a Lyapunov function for a non-linear SA with certain drift condition. Then, our theorem establishes finite-time bounds when this SA is driven by unbounded Markovian noise under suitable conditions. It serves as a black box tool to generalize sample guarantees on SA from i.i.d. or martingale difference case to potentially unbounded Markovian noise. The generality and the mild assumption of the setup enables broad applicability of our theorem. We illustrate its power by studying two more systems: (i) We improve upon the finite-time bounds of $Q$-learning by tightening the error bounds and also allowing for a larger class of behavior policies. (ii) We establish the first ever finite-time bounds for distributed stochastic optimization of high-dimensional smooth strongly convex function using cyclic block coordinate descent.
``Distribution shift'' is the main obstacle to the success of offline reinforcement learning. A learning policy may take actions beyond the behavior policy's knowledge, referred to as Out-of-Distribution (OOD) actions. The Q-values for these OOD actions can be easily overestimated. As a result, the learning policy is biased by using incorrect Q-value estimates. One common approach to avoid Q-value overestimation is to make a pessimistic adjustment. Our key idea is to penalize the Q-values of OOD actions associated with high uncertainty. In this work, we propose Q-Distribution Guided Q-Learning (QDQ), which applies a pessimistic adjustment to Q-values in OOD regions based on uncertainty estimation. This uncertainty measure relies on the conditional Q-value distribution, learned through a high-fidelity and efficient consistency model. Additionally, to prevent overly conservative estimates, we introduce an uncertainty-aware optimization objective for updating the Q-value function. The proposed QDQ demonstrates solid theoretical guarantees for the accuracy of Q-value distribution learning and uncertainty measurement, as well as the performance of the learning policy. QDQ consistently shows strong performance on the D4RL benchmark and achieves significant improvements across many tasks.
Power grid load scheduling is a critical task that ensures the balance between electricity generation and consumption while minimizing operational costs and maintaining grid stability. Traditional optimization methods often struggle with the dynamic and stochastic nature of power systems, especially when faced with renewable energy sources and fluctuating demand. This paper proposes a reinforcement learning (RL) approach using a Markov Decision Process (MDP) framework to address the challenges of dynamic load scheduling. The MDP is defined by a state space representing grid conditions, an action space covering control operations like generator adjustments and storage management, and a reward function balancing economic efficiency and system reliability. We investigate the application of various RL algorithms, from basic Q-Learning to more advanced Deep Q-Networks (DQN) and Actor-Critic methods, to determine optimal scheduling policies. The proposed approach is evaluated through a simulated power grid environment, demonstrating its potential to improve scheduling efficiency and adapt to variable demand patterns. Our results show that the RL-based method provides a robust and scalable solution for real-time load scheduling, contributing to the efficient management of modern power grids.