Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 44 for “"average reward"”.

  1. Learning Average Reward Irreducible Stochastic Games: Analysis and Applications

    … is used to extend the results of discounted reward stochastic games to average reward stochastic games. As a result, auxiliary matrix games are developed that have equivalent equilibrium points and values to a class of stochastic games that are irreducible and have average reward performance …

    usf Repository record for Learning Average Reward Irreducible Stochastic Games: Analysis and Applications (opens in a new tab)

  2. Policy-based average-reward and robust Markov decision processes and reinforcement learning

    Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-20 without embargo terms

    uiuc Repository record for Policy-based average-reward and robust Markov decision processes and reinforcement learning (opens in a new tab)

  3. Finite-sample Guarantees of Stochastic Approximation Algorithms in Average-reward Reinforcement Learning and Decentralized Stochastic Optimization

    … first provide the finite-sample guarantees for average-reward TD($\lambda$) with linear function approximation for policy evaluation, and tabular $Q$-learning in the synchronous setting for the control problem, addressing challenges posed by the lack of contraction in the average-reward setting. …

    gatech Repository record for Finite-sample Guarantees of Stochastic Approximation Algorithms in Average-reward Reinforcement Learning and Decentralized Stochastic Optimization (opens in a new tab)

  4. Efficient Algorithm for Learning and Generation with Reinforcement Learning

    … first comprehensive theoretical framework for average reward Markov potential games, establishing the smoothness of average reward functions and introducing three independent learning algorithms-policy gradient, proximal-Q, and natural policy gradient--with provable convergence to Nash …

    tamu Repository record for Efficient Algorithm for Learning and Generation with Reinforcement Learning (opens in a new tab) Embargoed until 2028-05-01

  5. Efficient Algorithm for Learning and Generation with Reinforcement Learning

    … first comprehensive theoretical framework for average reward Markov potential games, establishing the smoothness of average reward functions and introducing three independent learning algorithms-policy gradient, proximal-Q, and natural policy gradient--with provable convergence to Nash …

    tamu Repository record for Efficient Algorithm for Learning and Generation with Reinforcement Learning (opens in a new tab) Embargoed until 2028-05-01

  6. Towards Data Efficiency and Robustness of Reinforcement Learning

    … under uncertain environments. We adopt the average-reward criterion as an alternative reward measure for assessing long-term performance. Our research presents the first comprehensive study of robust MDPs under the average-reward criterion. We characterize the fundamental structure of robust …

    buffalo Repository record for Towards Data Efficiency and Robustness of Reinforcement Learning (opens in a new tab)

  7. Large Language Model Empowered Deep Reinforcement Learning to Support Prescriptive Maintenance

    … Compared with previous studies[1], the average reward of traditional deep reinforcement learning algorithms was 0.65, while after introducing large language models, the average reward steadily increased to 0.82. In addition, fluctuations in the early stages of learning were markedly …

    chalmers Repository record for Large Language Model Empowered Deep Reinforcement Learning to Support Prescriptive Maintenance (opens in a new tab)

  8. Integrated computational and network QOS in grid computing

    … known as Watkinsa?? Q(I>) and Semi-Markovian Average Reward Technique (SMART), to perform resource allocation on computing and network resources. We also explore two alternatives to resource allocation, provisioning and reservation. Our proposed solution was tested in simulation on a grid …

    nus Repository record for Integrated computational and network QOS in grid computing (opens in a new tab) Download file for Integrated computational and network QOS in grid computing (opens in a new tab)

  9. LINEAR PROGRAMMING APPROACHES FOR SEQUENTIAL DECISION MAKING: A COMPUTATIONAL PERSPECTIVE

    … second case, we consider an infinite-horizon, average reward MDP without known transition dynamics. Alternatively, the decision maker can access the transition information via a generative model. In this case, we adopted an inexact gradient method to solve the linear programming-based …

    nus Repository record for LINEAR PROGRAMMING APPROACHES FOR SEQUENTIAL DECISION MAKING: A COMPUTATIONAL PERSPECTIVE (opens in a new tab) Download file for LINEAR PROGRAMMING APPROACHES FOR SEQUENTIAL DECISION MAKING: A COMPUTATIONAL PERSPECTIVE (opens in a new tab)

  10. A Game-Theoretic Framework for Detecting Advanced Persistent Threats

    … the interactions between DIFT and APT as an average reward stochastic game to capture the long-term behavior of the APTs. We show the existence of an Average Reward Nash Equilibrium (ARNE) in DIFT-APT game. We propose a reinforcement learning algorithm, RL-ARNE, to learn an ARNE of DIFT-APT …

    washington Repository record for A Game-Theoretic Framework for Detecting Advanced Persistent Threats (opens in a new tab)

  11. Learning from Noisy and Delayed Rewards The Value of Reinforcement Learning to Defense Modeling and Simulation

    … algorithm that uses the exponentially weighted average reward as an action-value estimator is described. Empirical results indicate that this relatively straight-forward approach improves learning speed in both benchmark environments and in challenging applied settings. Applications of …

    nps Repository record for Learning from Noisy and Delayed Rewards The Value of Reinforcement Learning to Defense Modeling and Simulation (opens in a new tab)

  12. Safe Multi-Robot Planning Via Long-Run Averages

    … study investigated safe planning via long run average reward using MDPs. This thesis uses grid-world environments and builds on the Triple-QA framework [1]. Three approaches are evaluated: A single-agent baseline and two multi agent extensions, a trivial joint-state extension, and separate …

    chalmers Repository record for Safe Multi-Robot Planning Via Long-Run Averages (opens in a new tab)

  13. Testing Models of Context-Dependent Outcome Encoding in Reinforcement Learning

    … are centered around a running estimate of the average reward in each choice context, while the latter assumes that outcomes are compared to the minimum reward and then scaled by an estimate of the range of outcomes in each choice context. However, there are other computational mechanisms that …

    south-carolina Repository record for Testing Models of Context-Dependent Outcome Encoding in Reinforcement Learning (opens in a new tab)

  14. Joint lead time and price quotation : dynamic or static?

    … and formulate profit maximization models in an average reward criterion under both static and dynamic lead time and price quotations. Numerical analyses are presented to illustrate performances of both static and dynamic lead time and price quotation and thus find the threshold potential demand …

    lingnan Repository record for Joint lead time and price quotation : dynamic or static? (opens in a new tab)

  15. Energy Management System Design for Fuel Cell Vehicles

    … decision process based framework. We propose an average-reward approximator to estimate a long-term average cost instead of using a model to predict future power demand. Thus, the dependency between the system closed-loop performance and the model accuracy for predicting the future power demand …

    adelaide Repository record for Energy Management System Design for Fuel Cell Vehicles (opens in a new tab)

  16. Joint admission control and routing in IEEE 802.16-based mesh networks

    … classes can be prioritized by imposing different reward rates. We apply optimization techniques to obtain the optimal CAC policies. The optimality criterion is the long-run average reward. We demonstrate that the proposed scheme can the maximum revenue obtainable by the system under QoS …

    ubc Repository record for Joint admission control and routing in IEEE 802.16-based mesh networks (opens in a new tab) Download PDF for Joint admission control and routing in IEEE 802.16-based mesh networks (opens in a new tab)

  17. Neural mechanisms of reward-guided behavioural flexibility

    … such flexibility is guided by the prospect of reward. The first part of this thesis examines how reward prospect influences flexible updating of task rules, needed to switch between tasks. Chapter 2 investigates how sequential changes in reward prospect influence rule updating across three …

    oxford Repository record for Neural mechanisms of reward-guided behavioural flexibility (opens in a new tab)

  18. A Reinforcement Learning Approach To Obtain Treatment Strategies In Sequential Medical Decision Problems

    … algorithms of dynamic programming. An average reward reinforcement-learning algorithm is developed. The algorithm is applied on a sample problem of treating hereditary spherocytosis. The application demonstrates the ability of the proposed methodology to obtain effective treatment …

    usf Repository record for A Reinforcement Learning Approach To Obtain Treatment Strategies In Sequential Medical Decision Problems (opens in a new tab)

  19. Finite-sample analysis for distributed and neural reinforcement learning

    … continuous learning, it develops and analyzes average-reward TD-style algorithms that enforce a steady-state normalization using either a double-chain construction or a single-chain auxiliary recursion, proving finite-sample convergence rates that expose dependence on stepsizes, feature …

    bu Repository record for Finite-sample analysis for distributed and neural reinforcement learning (opens in a new tab)

Page 1 of 3