Search NASA⌕ Search

SEARCH · Search NASA

Results for “policy optimization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Decentralized Distributed Proximal Policy Optimization (DD-PPO) for High Performance Computing Scheduling on Multi-User Systems

Resource allocation in High Performance Computing (HPC) environments presents a complex and multifaceted challenge for job scheduling algorithms. Beyond the efficient allocation of system resources, schedulers must account for and optimize multiple performance metrics, including job wait time and system throughput. Traditional heuristic-based scheduling algorithms increasingly struggle and lack the efficiency needed to meet the demands and address the complexity and scale of modern HPC systems. Consequently, recent research efforts have focused on leveraging advancements in Artificial Intelligence (AI) and Deep Learning (DL), particularly Reinforcement Learning (RL), to develop more adaptable and intelligent scheduling strategies. Previous RL-based scheduling approaches have explored a range of algorithms, from Deep Q-Networks (DQN) to Proximal Policy Optimization (PPO), and more recently, hybrid methods that integrate Graph Neural Networks (GNNs) with RL techniques. However, a common limitation across these methods is their reliance on relatively small datasets, with few methods being evaluated using large-scale, multi-million-job trace datasets representative of real-world HPC workloads. Moreover, existing RL schedulers face scalability issues due to centralized policy updates, which hinder training efficiency and performance when applied to large datasets. This study introduces a novel RL-based scheduler utilizing Decentralized Distributed Proximal Policy Optimization (DD-PPO) algorithm, which supports large-scale distributed training across multiple workers without requiring parameter synchronization at every step. By eliminating reliance on centralized updates to a shared policy, the DD-PPO scheduler enhances scalability, training efficiency, and sample utilization. Experimental validation using a large real-world dataset containing over 11.5 million job traces collected from petascale HPC systems over six years assesses the influence of dataset scale on training effectiveness and compares DD-PPO performance to traditional and advanced scheduling approaches. The experimental results demonstrate improved scheduling performance in comparison to both heuristic-based schedulers and existing RL-based scheduling algorithms.

AI↗
NASA NTRS1975

Optimal policies for identification of stochastic linear systems

The problem of designing closed-loop policies for identification of multiinput-multioutput linear discrete-time systems with random time-varying parameters is considered in this paper using a Bayesian approach. A sensitivity index gives a measure of performance for the closed-loop laws. The computation of the optimal laws is shown to be nontrivial, an exercise in stochastic control, but open-loop, affine, and open-loop feedback optimal inputs are shown to yield tractable problems. Numerical examples are given. For time-invariant systems, the criterion considered is shown to be related to the trace of the information matrix associated with the system.

Lopez-Toledo, A. A.↗
NASA NTRS1982

Optimum equipment maintenance/replacement policy. Part 2: Markov decision approach

Dynamic programming was utilized as an alternative optimization technique to determine an optimal policy over a given time period. According to a joint effect of the probabilistic transition of states and the sequence of decision making, the optimal policy is sought such that a set of decisions optimizes the long-run expected average cost (or profit) per unit time. Provision of an alternative measure for the expected long-run total discounted costs is also considered. A computer program based on the concept of the Markov Decision Process was developed and tested. The program code listing, the statement of a sample problem, and the computed results are presented.

Charng, T.↗
DOE OSTI2025

Comprehensive assessment of deep reinforcement learning approaches for economic dispatch in nuclear-driven microgrids

As the electrical grid integrates more variable renewable energy sources such as wind and solar, the demand for distributed and flexible systems to address this increased variability becomes critical. Nuclear-driven microgrids provide a promising solution by offering stable generation to complement intermittent renewables, ensuring grid reliability and operating efficiency. This paper proposes a recurrent deep reinforcement learning framework for optimal economic dispatch in a nuclear-powered microgrid integrating renewable energy sources, small modular reactors, battery storage systems, and balance-of-plant dynamics. A three-agent control architecture is developed, where demand and renewable energy agents act as forecasters, and a reinforcement learning-based dispatch agent performs real-time energy allocation. A nonlinear programming formulation is first used to generate an optimal baseline for benchmarking. The proposed dispatch controller, based on Proximal Policy Optimization enhanced with Long Short-Term Memory networks, exploits temporal correlations in system dynamics by taking advantage of the time series used as inputs to improve policy robustness under uncertainty. Comparative analysis against established deep reinforcement learning methods, including Proximal Policy Optimization with a feedforward architecture, Soft Actor-Critic, and Twin Delayed Deep Deterministic Policy Gradient, demonstrates superior performance. Numerical results indicate that the proposed controller achieves a 0.39% cost reduction relative to the nonlinear programming benchmark and outperforms other learning-based methods by generating additional revenue of up to 0.35%. All reinforcement learning controllers compute dispatch actions in less than 0.3 s, resulting in a computational speedup of more than three orders of magnitude over the nonlinear programming baseline. The findings of this paper highlight their applicability for real-time operation and control in nuclear-integrated microgrids under volatile operating conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Safe Reinforcement Learning-Based Transient Stability Control for Islanded Microgrids With Topology Reconfiguration

This paper proposes a safe reinforcement learning (RL)-based transient stability emergency control (TSEC) method for islanded microgrids. RL requires extensive interaction with the environment to learn control strategies, hence, a data-driven approach is used as a substitute for time-consuming time-domain simulation calculations. Deep sigma point processes (DSPP), which is a Gaussian process model, is utilized to predict the normal distribution of transient stability of microgrids and to construct a transient stability chance constraint. Reward-constrained policy optimization (RCPO) can simultaneously achieve objective prediction, policy learning, and constraint cost coefficient update across multiple timescales. RCPO interacts with the DSPP-based microgrid environment through a multi-process parallel manner, greatly increasing the training speed. Case studies on a real islanded microgrid demonstrate that the proposed method can efficiently and quickly obtain the optimal emergency control strategy while adhering to all hard constraints.

14 SOLAR ENERGY↗
DOE OSTI2025

Black-box optimization of CT acquisition and reconstruction parameters: a reinforcement learning approach

Protocol optimization is critical in Computed Tomography (CT) for achieving desired diagnostic image quality while minimizing radiation dose. Due to the inter-effect of influencing CT parameters, traditional optimization methods rely on the testing of exhaustive combinations of these parameters. This poses a notable limitation due to the impracticality of exhaustive parameter testing. This study introduces a novel methodology leveraging Virtual Imaging Trials (VITs) and reinforcement learning to more efficiently optimize CT protocols. Computational phantoms with liver lesions were imaged using a validated CT simulator and reconstructed with a novel CT reconstruction Toolkit. The optimization parameter space included tube voltage, tube current, reconstruction kernel, slice thickness, and pixel size. The optimization process was done using a Proximal Policy Optimization (PPO) agent which was trained to maximize the Detectability Index (d’) of the liver lesion for each reconstructed image. Results showed that our reinforcement learning approach found the absolute maximum d’ across the test cases while requiring 79.7% fewer steps compared to an exhaustive search, demonstrating both accuracy and computational efficiency, offering a efficient and robust framework for CT protocol optimization. The flexibility of the proposed technique allows for use of varying image quality metrics as the objective metric to maximize for. Our findings highlight the advantages of combining VIT and reinforcement learning for CT protocol management.

Fenwick, David [Duke University Medical Center]↗
DOE OSTI2025

ARM-IRL: Adaptive Resilience Metric Quantification Using Inverse Reinforcement Learning

The resilience of safety-critical systems is gaining importance due to the rise in cyber and physical threats, especially within critical infrastructure. Traditional static resilience metrics may not capture dynamic system states, leading to inaccurate assessments and ineffective responses to cyber threats. This work aims to develop a data-driven, adaptive method for resilience metric learning. We propose a data-driven approach using inverse reinforcement learning (IRL) to learn a single, adaptive resilience metric. The method infers a reward function from expert control actions. Unlike previous approaches using static weights or fuzzy logic, this work applies adversarial inverse reinforcement learning (AIRL), training a generator and discriminator in parallel to learn the reward structure and derive an optimal policy. The proposed approach is evaluated on multiple scenarios: optimal communication network rerouting, power distribution network reconfiguration, and cyber–physical restoration of critical loads using the IEEE 123-bus system. The adaptive, learned resilience metric enables faster critical load restoration in comparison to conventional RL approaches.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

A Physics-Informed Reinforcement Learning Framework for Economic-Thermal Co-Optimization of Crypto Mining Data Centers: Preprint

The rapid expansion of cryptocurrency mining has created a new class of high-density data centers characterized by extreme thermal flux and high sensitivity to volatile economic markets. Traditional thermal management strategies, typically reliant on rule-based control, maintain static setpoints that fail to account for fluctuating electricity prices and cryptocurrency values - factors critical to mining profitability. To address this, we present a physics-informed reinforcement learning (PIRL) framework for economic-thermal co-optimization in crypto mining data centers. This framework consists of a proximal policy optimization (PPO) agent, a virtual testbed powered by high-fidelity physics-based models, and an interactive frontend dashboard. The PPO agent is trained using the virtual testbed and strict hardware safety limits. This physics-informed approach allows the agent to learn a stochastic policy that dynamically balances mining revenue against operational costs by co-optimizing HVAC cooling setpoints and IT computational hashrate. The simulation results demonstrate that the integrated framework achieved an 8.62% increase in net operational profit compared to traditional baseline strategies while strictly adhering to safety-critical temperature constraints (coolant supply temperature < 32 degrees C). This work provides a scalable template for the deployment of reinforcement learning in mission critical facilities where economic volatility and physical safety must be managed simultaneously.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

RLMolLM: Reinforcement Learning-Enhanced Language Model Framework for Inverse Molecular Design

Inverse molecular design faces significant challenges due to vast chemical space and complex property requirements. While language models show promise for molecular generation, they struggle with validity, multi-property optimization, and structural constraints. This work presents RLMolLM, a reinforcement learning framework combining Proximal Policy Optimization (PPO) with genetic algorithms to address these limitations. Our approach optimizes multiple user-specified properties including quantitative estimates of drug-likeness (QED), synthetic accessibility (SA), and ADMET (absorption, distribution, metabolism, excretion, and toxicity) endpoints without requiring complete model retraining, while maintaining capability for scaffold-constrained generation where specific substructures must be preserved. We outperform state-of-the-art methods for molecular optimization, achieving best QED scores across GDB13, Moses, and Zinc datasets with up to 31% improvement over previous methods while maintaining excellent validity, uniqueness, and novelty metrics. For simultaneous multi-property optimization, our framework achieves substantial improvements in ADMET properties including 4.5-fold reduction in hERG toxicity and enhanced Caco-2 permeability compared to Moses dataset. Under structural constraints, the framework significantly improves molecular validity while preserving scaffolds and effectively optimizing properties. In conclusion, this versatile solution advances pharmaceutical and materials molecular design through effective integration of reinforcement learning and genetic algorithms with multi-property optimization and scaffold preservation.

Genetic algorithms↗
NASA NTRS1978

Optimal analytic rendezvous using Clohessy-Wiltshire equations

The optimal solution time that minimizes the sum of the two applied impulses necessary to rendezvous is obtained analytically for the Clohessy-Wiltshire equations with a linear gravity model assumption. A plume impingement inequality constraint on the solution is examined, and an optimal policy is developed. Numerical tests are conducted to verify the analysis and to illustrate the optimal solution algorithm.

Jezewski, D. J.↗
NASA NTRS1979

An analytic approach to optimal rendezvous using Clohessy-Wiltshire equations

An analytic approach is used to obtain the optimal solution time that minimizes the sum of the two applied impulses necessary to rendezvous for the Clohessy-Wiltshire equations. A plume impingement inequality constraint on the solution is examined, and an optimal policy is developed. Numerical tests are conducted to verify the analysis and to illustrate the optimal solution algorithm.

Jezewski, D. J.↗
DOE OSTI2025

GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning

Large language models (LLMs) are increasingly adapted to downstream tasks via reinforcement learning (RL) methods like Group Relative Policy Optimization (GRPO), which often require thousands of rollouts to learn new tasks. We argue that the interpretable nature of language often provides a much richer learning medium for LLMs, compared to policy gradients derived from sparse, scalar rewards. To test this, we introduce GEPA (Genetic-Pareto), a prompt optimizer that thoroughly incorporates natural language reflection to learn high-level rules from trial and error. Given any AI system containing one or more LLM prompts, GEPA samples trajectories (e.g., reasoning, tool calls, and tool outputs) and reflects on them in natural language to diagnose problems, propose and test prompt updates, and combine complementary lessons from the Pareto frontier of its own attempts. As a result of GEPA's design, it can often turn even just a few rollouts into a large quality gain. Across six tasks, GEPA outperforms GRPO by 6% on average and by up to 20%, while using up to 35x fewer rollouts. GEPA also outperforms the leading prompt optimizer, MIPROv2, by over 10% (e.g., +12% accuracy on AIME-2025), and demonstrates promising results as an inference-time search strategy for code optimization. We release our code at https://github.com/gepa-ai/gepa.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1986

An optimal repartitioning decision policy

A central problem to parallel processing is the determination of an effective partitioning of workload to processors. The effectiveness of any given partition is dependent on the stochastic nature of the workload. The problem of determining when and if the stochastic behavior of the workload has changed enough to warrant the calculation of a new partition is treated. The problem is modeled as a Markov decision process, and an optimal decision policy is derived. Quantification of this policy is usually intractable. A heuristic policy which performs nearly optimally is investigated empirically. The results suggest that the detection of change is the predominant issue in this problem.

Nicol, D. M.↗
DOE OSTI2024

Design and optimization of a modular hydrogen-based integrated energy system to maximize revenue via nuclear-renewable sources

Here, this paper demonstrates a novel modular distributed framework that uses optimal energy-dispatching strategies to enable greater flexibility and profitability in nuclear-renewable integrated energy systems (NR-IES). Hydrogen is used as a commodity in this framework since its production can improve grid stability and system operational flexibility, decarbonize heavy industry, and create an additional revenue stream for electricity generators, particularly nuclear power plants with high operational expenses. The proposed solution addresses the challenges associated with merging multiple software and services from various domains by using functional mock-up units (FMU) to co-simulate diverse subsystems designed in various platforms. The tightly coupled integrated energy system (IES) is optimized to maximize revenue by utilizing the deep reinforcement learning (DRL) technique to make smart dispatching decisions based on variable electricity prices and the availability of renewable energy. Proximal policy optimization (PPO) algorithm is used in training and testing the DRL agent. Over a period of 120 days, the proposed hydrogen-based IES framework showed about 10% revenue boost compared to a non-hydrogen generating baseline IES while also providing an easily-adoptable framework which can help to improve the flexibility of future generation nuclear power plants.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
NASA NTRS1998

Markov Tracking for Agent Coordination

Partially observable Markov decision processes (POMDPs) axe an attractive representation for representing agent behavior, since they capture uncertainty in both the agent's state and its actions. However, finding an optimal policy for POMDPs in general is computationally difficult. In this paper we present Markov Tracking, a restricted problem of coordinating actions with an agent or process represented as a POMDP Because the actions coordinate with the agent rather than influence its behavior, the optimal solution to this problem can be computed locally and quickly. We also demonstrate the use of the technique on sequential POMDPs, which can be used to model a behavior that follows a linear, acyclic trajectory through a series of states. By imposing a "windowing" restriction that restricts the number of possible alternatives considered at any moment to a fixed size, a coordinating action can be calculated in constant time, making this amenable to coordination with complex agents.

Washington, Richard↗
DOE OSTI2026

Bridging Control and Deployment: A Cross-Layer Analysis of Scalable Building Cluster Control

Building cluster control has emerged as a promising approach for enabling flexible and coordinated operation of distributed building systems, yet its transition from pilot demonstrations to routine grid-interactive operation remains limited. This paper argues that this gap cannot be explained by control algorithms alone. Instead, it arises from interacting barriers in communication infrastructure, data and semantic interoperability, uncertainty management, stakeholder participation, market design, and policy support. Accordingly, the paper reviews both technical and non-technical barriers to building cluster control. Technical challenges include heterogeneous devices and protocols, communication latency and reliability, distributed decision-making, and uncertainty propagation across aggregated loads. Non-technical barriers include user participation, stakeholder coordination, incentive allocation, and data governance. Existing solution approaches are synthesized, including semantic interoperability frameworks, edge and hierarchical communication architectures, distributed and transactive control strategies, uncertainty-aware optimization, policy mechanisms, and market reforms. Based on this analysis, two research directions are identified: testing infrastructures that can evaluate control performance under realistic multi-building conditions, and abstraction methods that allow building clusters to interact with other energy sectors through standardized flexibility representations. Overall, the paper provides a structured review of how building cluster control can move from isolated demonstrations toward reproducible, market-compatible, and grid-relevant implementation.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

A Reinforcement Learning Approach to Augment Conventional PID Control in Nuclear Power Plant Transient Operation

The ability of nuclear reactors to operate their power conversion cycles more flexibly will enhance their value to energy grids with variable pricing. Current nuclear control systems are typically classical controllers that are often based on proportional-integral-derivative (PID) control. This paper presents a method of augmenting the existing PID control for difficult transient operations in nuclear power plants using a reinforcement learning–derived feedforward signal applied in real time. The agents, which are trained on a test thermal load-following problem, are designed to improve steam generator outlet temperature control for a range of fast load-following scenarios covering ramp rates from 9%/min to 15%/min. Several reinforcement learning algorithms were initially investigated for the training of the feedforward agents with deep Q-learning (DQN) and proximal policy optimization (PPO) networks, which were found to be the most promising. The DQN controllers utilize discrete actions, giving them a better disturbance rejection at steady state but inconsistent response to initial temperature deviations. In contrast, PPO-trained agents, which take continuous actions except for a dead zone around zero, were shown to have the best combination of high disturbance rejection at steady state and good tracking of the desired temperature value. The ability of the PPO agent was also examined, with the average time of decision making found to be on the order of 1 ms. The fault properties of the controller under the loss of the reinforcement learning agent feedforward signal were also examined. The controller showed strong performance in situations of “no-signal” faults. but was less good at handling “stuck-at” faults, where the feedforward signal remains at a set value. In both cases, however, the PID was able to successfully maintain stability, eventually returning the system to a steady state. It is hoped that this work will allow for the proposed control architecture to be examined for more difficult control problems such that it may eventually be used to adapt existing nuclear plants for more aggressive load-following on grids of the future.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗