SEARCH · Engineering Papers

Results for “Q-learning”

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.

SourceResearch connectionFrom yearThrough year

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

17 records

All sources NASA NTRS DOE OSTI

NASA NTRS1996

Fuzzy Q-Learning for Generalization of Reinforcement Learning

Fuzzy Q-Learning, introduced earlier by the author, is an extension of Q-Learning into fuzzy environments. GARIC is a methodology for fuzzy reinforcement learning. In this paper, we introduce GARIC-Q, a new method for doing incremental Dynamic Programming using a society of intelligent agents which are controlled at the top level by Fuzzy Q-Learning and at the local level, each agent learns and operates based on GARIC. GARIC-Q improves the speed and applicability of Fuzzy Q-Learning through generalization of input space by using fuzzy rules and bridges the gap between Q-Learning and rule based intelligent systems.

Berenji, Hamid R.↗

NASA NTRS2005

Quicker Q-Learning in Multi-Agent Systems

Multi-agent learning in Markov Decisions Problems is challenging because of the presence ot two credit assignment problems: 1) How to credit an action taken at time step t for rewards received at t' greater than t; and 2) How to credit an action taken by agent i considering the system reward is a function of the actions of all the agents. The first credit assignment problem is typically addressed with temporal difference methods such as Q-learning OK TD(lambda) The second credit assi,onment problem is typically addressed either by hand-crafting reward functions that assign proper credit to an agent, or by making certain independence assumptions about an agent's state-space and reward function. To address both credit assignment problems simultaneously, we propose the Q Updates with Immediate Counterfactual Rewards-learning (QUICR-learning) designed to improve both the convergence properties and performance of Q-learning in large multi-agent problems. Instead of assuming that an agent s value function can be made independent of other agents, this method suppresses the impact of other agents using counterfactual rewards. Results on multi-agent grid-world problems over multiple topologies show that QUICR-learning can achieve up to thirty fold improvements in performance over both conventional and local Q-learning in the largest tested systems.

Agogino, Adrian K.↗

NASA NTRS2006

QUICR-learning for Multi-Agent Coordination

Coordinating multiple agents that need to perform a sequence of actions to maximize a system level reward requires solving two distinct credit assignment problems. First, credit must be assigned for an action taken at time step t that results in a reward at time step t > t. Second, credit must be assigned for the contribution of agent i to the overall system performance. The first credit assignment problem is typically addressed with temporal difference methods such as Q-learning. The second credit assignment problem is typically addressed by creating custom reward functions. To address both credit assignment problems simultaneously, we propose the "Q Updates with Immediate Counterfactual Rewards-learning" (QUICR-learning) designed to improve both the convergence properties and performance of Q-learning in large multi-agent problems. QUICR-learning is based on previous work on single-time-step counterfactual rewards described by the collectives framework. Results on a traffic congestion problem shows that QUICR-learning is significantly better than a Q-learner using collectives-based (single-time-step counterfactual) rewards. In addition QUICR-learning provides significant gains over conventional and local Q-learning. Additional results on a multi-agent grid-world problem show that the improvements due to QUICR-learning are not domain specific and can provide up to a ten fold increase in performance over existing methods.

Agogino, Adrian K.↗

DOE OSTI2025

FIRE: A Failure-Adaptive RL Framework for Edge Computing Migrations

In edge computing, users' service profiles are migrated between edge servers due to user mobility. Reinforcement Learning (RL) frameworks have been proposed to do so, often trained on simulated data. However, existing RL frameworks overlook occasional server failures, which although rare, impact latency-sensitive applications like AR/VR and real- time obstacle detection. These rare failures, being not adequately represented in historical training data, pose a challenge for data-driven RL algorithms. We introduce FIRE, a framework that adapts to rare events by training a RL policy in an edge computing digital twin environment. We propose FIRE-ImRE, an importance sampling-based Q-learning algorithm, which samples rare events proportionally to their impact on the value function. FIRE considers delay, migration, failure, and backup placement costs across individual and shared service profiles. We prove FIRE-ImRE's boundedness and convergence to optimality. Next, we introduce novel deep Q-learning (FIRE-ImDQL) and actor critic (FIRE-ImACRE) versions of our algorithm to enhance scalability. Here, we extend our framework to accommodate users with varying risk tolerances of rare failure events. Through trace-driven experiments, we show that FIRE reduces edge computing costs compared to vanilla RL and the greedy baseline in the event of failures.

Edge computing↗

NASA NTRS2003

Reinforcement Learning for Weakly-Coupled MDPs and an Application to Planetary Rover Control

Weakly-coupled Markov decision processes can be decomposed into subprocesses that interact only through a small set of bottleneck states. We study a hierarchical reinforcement learning algorithm designed to take advantage of this particular type of decomposability. To test our algorithm, we use a decision-making problem faced by autonomous planetary rovers. In this problem, a Mars rover must decide which activities to perform and when to traverse between science sites in order to make the best use of its limited resources. In our experiments, the hierarchical algorithm performs better than Q-learning in the early stages of learning, but unlike Q-learning it converges to a suboptimal policy. This suggests that it may be advantageous to use the hierarchical algorithm when training time is limited.

Bernstein, Daniel S.↗

DOE OSTI2025

Swing Contract-Based Valuation for Distributed Energy Resources in Transactive Energy Systems: A Reinforcement Learning Approach

With the proliferation of distributed energy resources (DERs) and power grids with high fractions of renewable energy, market constructs are evolving to allow DERs to participate in multiple possible markets, at different levels of grid hierarchy. The effective participation of DERs in market environments is aided by swing contract-based pricing mechanisms, whereby DERs have a two-part compensation structure – one for their reservation/commitment and another for performancedriven ex-post payment for their actual mobilization during dispatch. In this paper, we propose a reinforcement learningbased (Q-learning) approach that allows a rational DER agent to select the market it wants to participate in within a composite market environment where individual markets are coordinated by possibly different actors. The proposed Q-learning framework aids DERs in their self-valuation by implicitly maximizing their own payoff through market participation, assuming a swing contract-based compensation structure. We complement our work through simulation-based investigations where factors affecting the DER decision making process, such as parametric uncertainties in market (and grid) environments, are studied.

Naqvi, Syed Ahsan Raza↗

DOE OSTI2024

Controlling optical-cavity locking using reinforcement learning

Abstract This study applies an effective methodology based on Reinforcement Learning to a control system. Using the Pound–Drever–Hall locking scheme, we match the wavelength of a controlled laser to the length of a Fabry-Pérot cavity such that the cavity length is an exact integer multiple of the laser wavelength. Typically, long-term drift of the cavity length and laser wavelength exceeds the dynamic range of this control if only the laser’s piezoelectric transducer is actuated, so the same error signal also controls the temperature of the laser crystal. In this work, we instead implement this feedback control grounded on Q-Learning. Our system learns in real-time, eschewing reliance on historical data, and exhibits adaptability to system variations post-training. This adaptive quality ensures continuous updates to the learning agent. This innovative approach maintains lock for eight days on average.

47 OTHER INSTRUMENTATION↗

NASA NTRSDate not supplied

Applying Machine Learning Tools for Runway Configuration Decision Support

Determining optimal runway configurations at airports, a responsibility assigned to air traffic controllers, is a challenging task. The decision-making process is intricate and involves consideration of many factors such as prevailing wind condition, convective weather, visibility, cloud ceilings, departure and arrival demand, traffic flow, equipment status, and other airport constraints. In a previous work, we developed a Runway Configuration Assistance tool using an offline reinforcement learning method called conservative Q-learning. In this paper, we evaluate and validate our Runway Configuration Assistance tool as a decision support for air traffic controllers. We validated our tool using three airports with differing levels of complexity: Charlotte Douglas International Airport, Denver International Airport, and Dallas Fort Worth International Airport. We quantified the performance of the Runway Configuration Assistance tool based on (1) agreement with historical air traffic controller decisions and (2) violation of decisions that would be obvious to subject-matter experts. Our tool showed promising results in both performance metrics for the three airports, despite the complexities in the runway configuration decision-making process. We also discuss challenges in using machine learning in general to aid air traffic management and identify deployment considerations for the Runway Configuration Assistance tool.

Runway Configuration Management↗

NASA NTRSDate not supplied

Machine Learning Tools for Runway Configuration Decision Support

Runway Configuration Management↗

DOE OSTI2026

Cascading economic losses from port disruptions under capacity constrained multimodal freight networks

This study quantifies how throughput disruptions at major seaports cascade through capacity-constrained multimodal freight networks and interregional production systems. We couple an agent-based model (ABM) multimodal freight simulation that resolves rerouting, terminal queueing, and inventory drawdown under binding modal and facility capacities with a multiregional output loss input-output (MRIIM) model that propagates realized delivery shortfalls across regions and sectors. The framework is demonstrated for the Port of Los Angeles using Freight Analysis Framework flows and Bureau of Economic Analysis input-output accounts and is evaluated over a 52-week horizon under deterministic sector targeted shocks and stochastic disruption realizations with uncertain severity and duration. Results indicate nonlinear amplification: realized national losses concentrate in manufacturing and transportation/warehousing even when exogenous port shocks are dispersed, suggesting that congestion spillback and limited short-run substitution can dominate the initial shock allocation. We further evaluate a tabular reinforcement-learning (Q-learning) intervention layer that selects among a small set of implementable system level levers (truck-to-rail and truck-to-barge shift settings) without overriding shipper routing, finding that such interventions reduce total losses for moderate disruptions but yield diminishing returns once substitute modes approach capacity. By linking operational freight behavior to system wide impacts under uncertainty, the proposed ABM-MRIIM pipeline provides a reusable workflow for port disruption stress testing, identification of structurally critical sectors/corridors, and evaluation of resilience interventions under realistic capacity limits.

42 ENGINEERING↗

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↗

DOE OSTI2024

Short-Term Probabilistic Solar Forecasting via Reinforcement Learning over ECMWF

In this paper, we present an innovative reinforcement learning approach for short-term solar forecasting, leveraging data from the European Centre for Medium-Range Weather Forecasts (ECMWF). The methodology begins with the application of the System Advisor Model (SAM) to transform various ECMWF numerical weather prediction members into predictive photovoltaic power generation. To enhance the precision of deterministic forecasting, we introduce a dynamic model selection algorithm based on Q-learning. This algorithm dynamically identifies and utilizes the most accurate ensemble member for forecasting purposes. Furthermore, we employ a support vector regression surrogate model with a Gaussian distribution to generate probabilistic forecasts, providing a holistic view of solar energy generation uncertainty. To expedite the training process and make it more practical for real-world applications, we integrate a rolling update workflow. This innovative workflow reduces the training period from months to a mere 19 days, making our method highly efficient. Numerical results of the case study show that in comparison to benchmark models, the proposed method improves the deterministic and probabilistic solar forecasting accuracy by up to 40.84% and 48.42%, respectively.

ensemble forecasting↗

NASA NTRS1991

On the integration of reinforcement learning and approximate reasoning for control

The author discusses the importance of strengthening the knowledge representation characteristic of reinforcement learning techniques using methods such as approximate reasoning. The ARIC (approximate reasoning-based intelligent control) architecture is an example of such a hybrid approach in which the fuzzy control rules are modified (fine-tuned) using reinforcement learning. ARIC also demonstrates that it is possible to start with an approximately correct control knowledge base and learn to refine this knowledge through further experience. On the other hand, techniques such as the TD (temporal difference) algorithm and Q-learning establish stronger theoretical foundations for their use in adaptive control and also in stability analysis of hybrid reinforcement learning and approximate reasoning-based controllers.

Berenji, Hamid R.↗

NASA NTRS2004

Unifying Temporal and Structural Credit Assignment Problems

Single-agent reinforcement learners in time-extended domains and multi-agent systems share a common dilemma known as the credit assignment problem. Multi-agent systems have the structural credit assignment problem of determining the contributions of a particular agent to a common task. Instead, time-extended single-agent systems have the temporal credit assignment problem of determining the contribution of a particular action to the quality of the full sequence of actions. Traditionally these two problems are considered different and are handled in separate ways. In this article we show how these two forms of the credit assignment problem are equivalent. In this unified frame-work, a single-agent Markov decision process can be broken down into a single-time-step multi-agent process. Furthermore we show that Monte-Carlo estimation or Q-learning (depending on whether the values of resulting actions in the episode are known at the time of learning) are equivalent to different agent utility functions in a multi-agent system. This equivalence shows how an often neglected issue in multi-agent systems is equivalent to a well-known deficiency in multi-time-step learning and lays the basis for solving time-extended multi-agent problems, where both credit assignment problems are present.

Agogino, Adrian K.↗

NASA NTRSDate not supplied

Deep Reinforcement Learning based Routing in an Air-to-Air Ad-hoc Network

This paper studies the Multiple Sources and Multiple Destinations (MSMD) routing problem in a dynamic Air-to-Air Ad-hoc Network (AAAN). We consider a spectrum limited scenario where multiple links have to share the same frequency channel so that co-channel interference becomes inevitable. As a result, routing decisions and spectrum access are coupled and must be jointly considered. This paper proposes a deep Q-learning based algorithm to find an optimal routing and channel selection strategy that minimizes the end-to-end communication delay. Specifically, under the assumption that only local information is available to every node, the Deep Q-Network (DQN) is trained offline to learn the optimal routing and channel selection strategy. After the trained DQN is implemented in every node, multiple relay nodes can simultaneously determine their next-hop relay and channel selections in real-time. Simulation results demonstrate the efficacy of our proposed algorithm.

AAAN↗

NASA NTRSDate not supplied

Deep Reinforcement Learning based Routing in an Air-to-Air Ad-hoc Network

AAAN↗

DOE OSTI2024

Reinforcement Learning Control for Enhancing Marine Hydrokinetic Turbine Energy Generation

This paper proposes a reinforcement learning-based method to maximize power generation for a direct-drive marine hydrokinetic turbine. A high levelized cost of energy (LCOE) is preventative in the widespread adoption of many marine energy conversion technologies. A straightforward way to reduce LCOE is to increase conversion efficiency and ensure maximum energy generation. The proposed method utilizes a damping control methodology, varying applied generator torque via a linear relationship between the applied damping coefficient and rotor speed. A state-action-reward-state-action (SARSA) algorithm has been used to learn the optimal control action for a given flow velocity. The proposed SARSA methodology uses Gaussian radial basis functions to create a three-dimensional surface to estimate the relationship between damping coefficient, incoming flow velocity, and coefficient of power (C p ). Here, the SARSA algorithm was compared against a baseline optimal tip speed ratio controller over a year-long flow velocity case profile while considering the effects of biofouling on the turbine system, where the proposed RL method generated 0.92% more energy than the baseline.

Damp↗