Search NASA⌕ Search

SEARCH · Search NASA

Results for “, policy gradient”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Deploying Traffic Smoothing Cruise Controllers Learned from Trajectory Data

Autonomous vehicle-based traffic smoothing con- trollers are often not transferred to real-world use due to challenges in calibrating many-agent traffic simulators. We show a pipeline to sidestep such calibration issues by collecting trajectory data and learning controllers directly from trajectory data that are then deployed zero-shot onto the highway. We construct a dataset of 772.3 kilometers of recorded drives on the I-24. We then construct a simple simulator using the recorded drives as the lead vehicle in front of a simulated platoon consisting of one autonomous vehicle and five human followers. Using policy-gradient methods with an asymmetric critic to learn the controller, we show that we are able to improve average MPG by 11% in simulation on congested trajectories. We deploy this controller to a mixed platoon of 4 autonomous Toyota RAV-4’s and 7 human drivers in a validation experiment and demonstrate that the expected time-gap of the controller is maintained in the real world test. Finally, we release the driving dataset [1], the simulator, and the trained controller at https://github.com/nathanlct/trajectory-training-icra.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Data-Driven Distribution System Coordinated PV Inverter Control Using Deep Reinforcement Learning

The deployment of distributed solar photovoltaic (PV) systems has increased consistently over the past decades. High penetrations of PVs could cause a series of adverse grid impacts, such as voltage violations. The recent development of smart inverter technologies rises the incentives of developing PV control solutions that regulate the inverter output power and seeking the optimization on system operational objectives. This paper proposes a data-driven control solution based on deep reinforcement learning (DRL) to optimize PV inverters for voltage regulation. The proposed solution can minimize PV real power curtailment while maintaining network voltage at an acceptable range. Comparison results between the proposed DRL control algorithms with deep deterministic policy gradient (DDPG) and volt-var control on a real feeder in west Colorado highlight the advantage of the proposed framework in controlling the system voltage while minimizing the PV real power curtailment.

deep reinforcement learning↗
DOE OSTI2024

Federated Deep Reinforcement Learning for Decentralized VVO of BTM DERs

The future of grid control requires a hybrid approach combining centralized and decentralized methods to fully utilize the potential of smart edge devices with artificial intelligence (AI) capabilities. This paper aims to develop and evaluate a federated deep reinforcement learning (FDRL) framework for decentralized adaptive volt-var optimization (VVO) of behind-the-meter (BTM) distributed energy resources (DERs). First, this paper models a single deep reinforcement learning (DRL) agent using the Markov Decision Process (MDP) framework for decentralized adaptive VVO of BTM DERs. Two DRL algorithms, soft actor-critic (SAC) and twin-delayed deep deterministic policy gradient (TD3), are compared for their effectiveness in optimizing VVO. Results show that TD3 outperforms SAC, achieving a 71.3% improvement in mean reward. Finally, the DRL agent is deployed within the FDRL framework, using the Flower platform, to enhance learning, provide adaptive control, and ensure data privacy for BTM DERs.

Ravi, Abhijith↗
DOE OSTI2022

Post-Disaster Microgrid Formation for Enhanced Distribution System Resilience

This paper proposes a deep reinforcement learning (DRL) based approach for post-disaster critical load restoration in active distribution systems to form microgrids through network reconfiguration to minimize critical load curtailments. Distribution networks are represented as graph networks, and optimal network configurations with microgrids are obtained by searching for the optimal spanning forest. The constraints to the research question being explored are the radial topology and power balance. Unlike existing analytical and population-based approaches, which necessitate the repetition of entire analyses and computation for each outage scenario to find the optimal spanning forest, the proposed approach, once properly trained, can quickly determine the optimal, or near-optimal, spanning forest even when outage scenarios change. When multiple lines fail in the system, the proposed approach forms microgrids with distributed energy resources in active distribution systems to reduce critical load curtailment. The proposed DRL-based model learns the action-value function using the REINFORCE algorithm, which is a model-free reinforcement learning technique based on stochastic policy gradients. A case study was conducted on a 33-node distribution test system, demonstrating the effectiveness of the proposed approach for post-disaster critical load restoration.

active distribution systems↗
DOE OSTI2024

Optimal Control of SOEC-Based Hydrogen Production Systems for Demand Response Using Deep Reinforcement Learning in Smart Grids

Solid oxide electrolysis cell (SOEC) hydrogen production technology can range in size from small, appliance-size equipment to large-scale, central production facilities that can be tied directly to renewable or non-greenhouse-gas-emitting forms of electricity production, making it an ideal resource for demand response (DR). The SOEC hydrogen production system is a complex integrated system that encompasses fluid dynamics, electrical dynamics, and electrochemical and thermal dynamics, all of which involve non-linearity and non-convexity. Proper control of the SOEC hydrogen production system is crucial to enable its participation in the DR program. Here, to overcome the difficulty of designing an explicit control law for such nonlinear systems with nonconvex optimization features in DR applications, deep reinforcement learning (DRL) is explored to achieve the optimal control of the SOEC system for DR participation. Specifically, a twin delayed deterministic policy gradient (TD3) control framework is applied to achieve optimal response performance during DR events by considering power tracking error and hydrogen production efficiency with a suitable reward function. Two case studies with grid connections for tracking different DR commands were investigated. The first case study involved operating conditions reaching the boundaries, while the second involved operating conditions within the boundaries. The results showed that the proposed DRL-based control for SOEC can track the DR signal in a timely manner while maintaining high energy efficiency.

08 HYDROGEN↗
DOE OSTI2024

From Sim to Real: A Pipeline for Training and Deploying Traffic Smoothing Cruise Controllers

Designing and validating controllers for connected and automated vehicles to enhance traffic flow presents significant challenges, from the complexity of replicating real-world stop-and-go traffic dynamics in simulation, to the intricacies involved in transitioning from simulation to actual deployment. In this work, we present a full pipeline from data collection to controller deployment. Specifically, we collect 772 km of driving data from the I-24 in Tennessee, and use it to build a one-lane simulator, placing simulated vehicles behind real-world trajectories. Using policy-gradient methods with an asymmetric critic, we improve fuel efficiency by over 10% when simulating congested scenarios. Our comprehensive approach includes reinforcement learning for controller training, software verification, hardware validation and setup, and navigating various sim-to-real challenges. Furthermore, we analyze the controller's behavior and wave-smoothing properties, and deploy it on four Toyota Rav4’s in a real-world validation experiment on the I-24. Lastly, we release the driving dataset, the simulator and the trained controller, to enable future benchmarking and controller design.

42 ENGINEERING↗
DOE OSTI2020

Automated Control of Transactive HVACs in Energy Distribution Systems

Heating, Ventilation, and Air Conditioning (HVAC) systems contribute significantly to a building’s energy consumption. In the recent years, there is an increased interest in developing transactive approaches which could enable automated and flexible scheduling of HVAC systems based on the customer demand and the electricity prices decided by the suppliers. Flexible and automated scheduling of the HVAC systems make it a prime source for participation in residential demand response or transactive energy systems. Therefore, it is of significant interest to identify an optimal strategy to control the HVAC systems. Here, reducing the energy cost while keeping the comfort level acceptable to the users, we argue that such a control strategy should consider both the energy cost and user comfort simultaneously. Accordingly, we develop the control strategy through the solution of an optimization problem that balances between the energy cost and consumer’s dissatisfaction. This optimization enables us to solve a decision-making problem through first price prediction and then choosing HVAC temperature settings throughout the day based on the predicted price, history of the price and HVAC settings, and outside temperature. More specifically, we formulate the control design as a Markov decision process (MDP) using deep neural networks and use Deep Deterministic Policy Gradients (DDPG)-based deep reinforcement learning algorithm to find the optimal control strategy for HVAC systems that balances between electricity cost and user comfort.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Multirobot Collaborative Pursuit Target Robot by Improved MADDPG

Policy formulation is one of the main problems in multirobot systems, especially in multirobot pursuit-evasion scenarios, where both sparse rewards and random environment changes bring great difficulties to find better strategy. Existing multirobot decision-making methods mostly use environmental rewards to promote robots to complete the target task that cannot achieve good results. This paper proposes a multirobot pursuit method based on improved multiagent deep deterministic policy gradient (MADDPG), which solves the problem of sparse rewards in multirobot pursuit-evasion scenarios by combining the intrinsic reward and the external environment. The state similarity module based on the threshold constraint is as a part of the intrinsic reward signal output by the intrinsic curiosity module, which is used to balance overexploration and insufficient exploration, so that the agent can use the intrinsic reward more effectively to learn better strategies. The simulation experiment results show that the proposed method can improve the reward value of robots and the success rate of the pursuit task significantly. The intuitive change is obviously reflected in the real-time distance between the pursuer and the escapee, the pursuer using the improved algorithm for training can get closer to the escapee more quickly, and the average following distance also decreases.

Zhou, Xiao↗
DOE OSTI2022

A Deep Reinforcement Learning-based Reserve Optimization in Active Distribution Systems for Tertiary Frequency Regulation

Federal Energy Regulatory Commission (FERC)Orders 841 and 2222 have recommended that distributed energy resources (DERs) should participate in energy and reserve markets; therefore, a mechanism needs to be developed to facilitate DERs’ participation at the distribution level. Although the available reserve from a single distribution system may not be sufficient for tertiary frequency regulation, stacked and coordinated contributions from several distribution systems can enable them participate in tertiary frequency regulation at scale. This paper proposes a deep reinforcement learning (DRL)-based approach for optimization of requested aggregated reserves by system operators among the clusters of DERs. The co-optimization of cost of reserve, distribution network loss, and voltage regulation of the feeders are considered while optimizing the reserves among participating DERs. The proposed framework adopts deep deterministic policy gradient (DDPG), which is an algorithm based on an actor-critic method. The effectiveness of the proposed method for allocating reserves among DERs is demonstrated through case studies on a modified IEEE 34-node distribution system.

deep reinforcement learning, distributed energy re↗
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↗
DOE OSTI2021

Spatial Graph Attention and Curiosity-driven Policy for Antiviral Drug Discovery

We developed Distilled Graph Attention Policy Networks (DGAPNs), a curiosity-driven reinforcement learning model to generate novel graph-structured chemical representations that optimize user-defined objectives by efficiently navigating a physically constrained domain. Here, the framework is examined on the task of generating molecules that are designed to bind, noncovalently, to functional sites of SARS-CoV-2 proteins. We present a spatial Graph Attention Network (sGAT) that leverages self-attention over both node and edge attributes as well as encoding spatial structure -- this capability is of considerable interest in areas such as molecular and synthetic biology and drug discovery. An attentional policy network is then introduced to learn decision rules for a dynamic, fragment-based chemical environment, and state-of-the-art policy gradient techniques are employed to train the network with enhanced stability. Exploration is efficiently encouraged by incorporating innovation reward bonuses learned and proposed by random network distillation. In experiments, our framework achieved outstanding results compared to state-of-the-art algorithms, while increasing the diversity of proposed molecules and reducing the complexity of paths to chemical synthesis.

Wu, Yulun↗
DOE OSTI2022

Spatial Graph Attention and Curiosity-driven Policy for Antiviral Drug Discovery

We developed Distilled Graph Attention Policy Network (DGAPN), a reinforcement learning model to generate novel graph-structured chemical representations that optimize user-defined objectives by efficiently navigating a physically constrained domain. The framework is examined on the task of generating molecules that are designed to bind, noncovalently, to functional sites of SARS-CoV-2 proteins. We present a spatial Graph Attention (sGAT) mechanism that leverages self-attention over both node and edge attributes as well as encoding the spatial structure --- this capability is of considerable interest in synthetic biology and drug discovery. An attentional policy network is introduced to learn the decision rules for a dynamic, fragment-based chemical environment, and state-of-the-art policy gradient techniques are employed to train the network with stability. Exploration is driven by the stochasticity of the action space design and the innovation reward bonuses learned and proposed by random network distillation. In experiments, our framework achieved outstanding results compared to state-of-the-art algorithms, while reducing the complexity of paths to chemical synthesis.

Wu, Yulun↗
DOE OSTI2022

Spatial Graph Attention and Curiosity-driven Policy for Antiviral Drug Discovery

We developed Distilled Graph Attention Policy Networks (DGAPNs), a curiosity-driven reinforcement learning model to generate novel graph-structured chemical representations that optimize user-defined objectives by efficiently navigating a physically constrained domain. The framework is examined on the task of generating molecules that are designed to bind, noncovalently, to functional sites of SARS-CoV-2 proteins. We present a spatial Graph Attention Network (sGAT) that leverages self-attention over both node and edge attributes as well as encoding spatial structure -- this capability is of considerable interest in areas such as molecular and synthetic biology and drug discovery. An attentional policy network is then introduced to learn decision rules for a dynamic, fragment-based chemical environment, and state-of-the-art policy gradient techniques are employed to train the network with enhanced stability. Exploration is efficiently encouraged by incorporating innovation reward bonuses learned and proposed by random network distillation. In experiments, our framework achieved outstanding results compared to state-of-the-art algorithms, while increasing the diversity of proposed molecules and reducing the complexity of paths to chemical synthesis.

Wu, Yulun↗
DOE OSTI2023

A Novel Multi-Agent Deep Reinforcement Learning-enabled Distributed Power Allocation Scheme for mmWave Cellular Networks

We consider the power allocation problem over shared spectrum for millimeter-Wave (mmWave) cellular downlink. Existing approaches usually find sub-optimal solutions by solving a non-convex optimization which leads to scalability issues due to centralized control. Therefore, distributed and adaptive approaches are desirable. Recently, model-free Deep Reinforcement Learning (DRL) has achieved success in such wireless resource management tasks. By modeling the radio environment as a Markov Decision Process (MDP) with the base stations (BSs) being the agents, power allocation can be automated at the agent level with comparable throughput performance to conventional centralized schemes. The multi-agent setting presents new challenges as the radio environment is impacted by the joint actions of the agents and is no longer stationary from any individual agent’s perspective. Existing literature bypasses this non-stationarity violation by ignoring it which may cause performance degradation. To tackle this issue, we propose a distributed continuous power allocation scheme based on a modified version of multi-agent Deep Deterministic Policy Gradient (MADDPG) that is tailored for the distributed multiple-agent setting. The proposed scheme employs a centralized-training distributed-execution framework where Q-functions are trained over subsets of BSs while each BS determines its transmit power based only on its own local observation. It admits constant per-BS communication and computation complexity and is thus scalable to large networks. Numerical evaluation shows that the proposed scheme adapts well to a wide range of interference conditions and can achieve comparable or better performance than several state-of-the-art non-learning approaches.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2022

SPATIAL GRAPH ATTENTION AND CURIOSITY-DRIVEN POLICY FOR ANTIVIRAL DRUG DISCOVERY

We developed Distilled Graph Attention Policy Network (DGAPN), a reinforcement learning model to generate novel graph-structured chemical representations that optimize user-defined objectives by efficiently navigating a physically constrained domain. The framework is examined on the task of generating molecules that are designed to bind, noncovalently, to functional sites of SARS-CoV-2 proteins. We present a spatial Graph Attention (sGAT) mechanism that leverages self-attention over both node and edge attributes as well as encoding the spatial structure - this capability is of considerable interest in synthetic biology and drug discovery. An attentional policy network is introduced to learn the decision rules for a dynamic, fragment-based chemical environment, and state-of-the-art policy gradient techniques are employed to train the network with stability. Exploration is driven by the stochasticity of the action space design and the innovation reward bonuses learned and proposed by random network distillation. In experiments, our framework achieved outstanding results compared to state-of-the-art algorithms, while reducing the complexity of paths to chemical synthesis.

Wu, Y↗
DOE OSTI2023

Hierarchical Flexibility Offering Strategy for Integrated Hybrid Resources in Real-time Energy Markets

This paper proposes a hierarchical model for determining the energy flexibility offering strategy of integrated hybrid resources (IHRs) in power distribution systems to participate in real-time energy markets. The proposed model utilizes the scalability, fast response time, and uncertainty observation of deep reinforcement learning (DRL) to overcome the scalability issue of operating numerous flexible resources and deliverability of energy flexibility to the real-time markets in the presence of the network constraints. To that end, the power distribution system is divided into multiple IHRs, where different types of flexible loads, energy storage systems, and solar plants with controllable inverters are operated through local IHR controllers, trained by deep deterministic policy gradient (DDPG) algorithm. Active power request and reactive power capacity of IHRs are then transmitted to a central flexibility controller, where a quadratic optimization model ensures the deliverability of the energy flexibility to the real-time energy market by satisfying the distribution network constraints. The proposed model is implemented on the 123-bus test power distribution system, demonstrating the capability of DRL-based hierarchical model for scalable operation of IHRs in order to offer deliverable energy flexibility to the real-time energy market.

Majidi, Majid↗
DOE OSTI2024

Optimal Coordination of Electric Vehicles for Grid Services using Deep Reinforcement Learning

Recent research has shown the effectiveness of reinforcement learning (RL) in coordinating electric vehicles (EVs) with vehicle-to-grid capabilities for grid services. However, many of these studies rely on lookup table and deep Q-network techniques, which can be impractical when dealing with continuous states and actions. In addition, existing RL designs inadequately account for battery aging effects, EV user satisfaction, uncertain departure and arrival time, and trip distance, which may compromise effective coordination. This paper aims to bridge these gaps by developing an innovative deep deterministic policy gradient-based RL framework for optimal coordination of EVs. Case studies were carried out using a test system with 100 EVs, and numerical analysis results showed that the proposed RL framework can effectively coordinate EVs to maximize economic benefits and user satisfaction while ensuring the expected battery lifespan.

Das, Avijit↗