Existing Reinforcement Learning Methods Research Articles

Autonomous UAV Navigation with Adaptive Control Based on Deep Reinforcement Learning

Unmanned aerial vehicle (UAV) navigation plays a crucial role in its ability to perform autonomous missions in complex environments. Most of the existing reinforcement learning methods to solve the UAV navigation problem fix the flight altitude and velocity, which largely reduces the difficulty of the algorithm. But the methods without adaptive control are not suitable in low-altitude environments with complex situations, generally suffering from weak obstacle avoidance. Some UAV navigation studies with adaptive flight only have weak obstacle avoidance capabilities. To address the problem of UAV navigation in low-altitude environments, we construct autonomous UAV navigation in 3D environments with adaptive control as a Markov decision process and propose a deep reinforcement learning algorithm. To solve the problem of weak obstacle avoidance, we creatively propose the guide attention method to make a UAV’s decision focus shift between the navigation task and obstacle avoidance task according to changes in the obstacle. We raise a novel velocity-constrained loss function and add it to the original actor loss to improve the UAV’s velocity control capability. Simulation experiment results demonstrate that our algorithm outperforms some of the state-of-the-art deep reinforcement learning algorithms performing UAV navigation tasks in a 3D environment and has outstanding performance in algorithm effectiveness, with the average reward increasing by 9.35, the success rate of navigation tasks increasing by 14%, and the collision rate decreasing by 14%.

32‐3: Heterogeneous Resource Constrained Reinforcement Learning Photolithography Scheduler with Heterogeneous Graph Attention Network

In mass manufacturing of OLED displays, photolithography process is a critical bottleneck in production capacity. As a result, effective scheduling of the photolithography process is crucial to the overall throughput, productivity, and efficiency of the production. The processing of OLED panels depends upon heterogeneous resources. Aside from that, the complexity, stochastic nature, and highly dynamic characteristic of the processpresent considerable challenges. To address these, we propose a new framework employing a heterogeneous graph neural networkbased Rainbow algorithm. This framework optimizes the schedule used in photolithography process. Considering the constraints on machines, masks, and the stochastic nature of arrival processes, we optimize for maximizing productivity while minimizing the associated costs. We model the interactions among product lot steps, machines, and masks as a heterogeneous graph. This graph encodes the information of lot steps, machines, and masks into distinctive nodes. We then implement a Graph Attention Networkbased architecture for deep representation learning. This transforms state information into node embeddings for each step of each product lot, facilitating decision‐making. Reinforcement learning agents leverage these embeddings to prioritize products via a parameterized Q‐function. Our extensive experiments demonstrate the superior performance of our approach across multiple evaluation metrics in both static and dynamic environment benchmarks, with a higher winning rate and scheduling reward return when compared to existing reinforcement learning methods.

Combining Reinforcement Learning and Tensor Networks, with an Application to Dynamical Large Deviations.

We present a framework to integrate tensor network (TN) methods with reinforcement learning (RL) for solving dynamical optimization tasks. We consider the RL actor-critic method, a model-free approach for solving RL problems, and introduce TNs as the approximators for its policy and value functions. Our "actor-critic with tensor networks" (ACTeN) method is especially well suited to problems with large and factorizable state and action spaces. As an illustration of the applicability of ACTeN we solve the exponentially hard task of sampling rare trajectories in two paradigmatic stochastic models, the East model of glasses and the asymmetric simple exclusion process, the latter being particularly challenging to other methods due to the absence of detailed balance. With substantial potential for further integration with the vast array of existing RL methods, the approach introduced here is promising both for applications in physics and to multi-agent RL problems more generally.

Multimodality Driven Impedance-Based Sim2Real Transfer Learning for Robotic Multiple Peg-in-Hole Assembly.

Robotic rigid contact-rich manipulation in an unstructured dynamic environment requires an effective resolution for smart manufacturing. As the most common use case for the intelligence industry, a lot of studies based on reinforcement learning (RL) algorithms have been conducted to improve the performances of single peg-in-hole assembly. However, existing RL methods are difficult to apply to multiple peg-in-hole issues due to more complicated geometric and physical constraints. In addition, previously limited solutions for multiple peg-in-hole assembly are hard to transfer into real industrial scenarios flexibly. To effectively address these issues, this work designs a novel and more challenging multiple peg-in-hole assembly setup by using the advantage of the Industrial Metaverse. We propose a detailed solution scheme to solve this task. Specifically, multiple modalities, including vision, proprioception, and force/torque, are learned as compact representations to account for the complexity and uncertainties and improve the sample efficiency. Furthermore, RL is used in the simulation to train the policy, and the learned policy is transferred to the real world without extra exploration. Domain randomization and impedance control are embedded into the policy to narrow the gap between simulation and reality. Evaluation results demonstrate the effectiveness of the proposed solution, showcasing successful multiple peg-in-hole assembly and generalization across different object shapes in real-world scenarios.

Reinforcement learning-based SDN routing scheme empowered by causality detection and GNN.

In recent years, with the rapid development of network applications and the increasing demand for high-quality network service, quality-of-service (QoS) routing has emerged as a critical network technology. The application of machine learning techniques, particularly reinforcement learning and graph neural network, has garnered significant attention in addressing this problem. However, existing reinforcement learning methods lack research on the causal impact of agent actions on the interactive environment, and graph neural network fail to effectively represent link features, which are pivotal for routing optimization. Therefore, this study quantifies the causal influence between the intelligent agent and the interactive environment based on causal inference techniques, aiming to guide the intelligent agent in improving the efficiency of exploring the action space. Simultaneously, graph neural network is employed to embed node and link features, and a reward function is designed that comprehensively considers network performance metrics and causality relevance. A centralized reinforcement learning method is proposed to effectively achieve QoS-aware routing in Software-Defined Networking (SDN). Finally, experiments are conducted in a network simulation environment, and metrics such as packet loss, delay, and throughput all outperform the baseline.

NondBREM: Nondeterministic Offline Reinforcement Learning for Large-Scale Order Dispatching

One of the most important tasks in ride-hailing is order dispatching, i.e., assigning unserved orders to available drivers. Recent order dispatching has achieved a significant improvement due to the advance of reinforcement learning, which has been approved to be able to effectively address sequential decision-making problems like order dispatching. However, most existing reinforcement learning methods require agents to learn the optimal policy by interacting with environments online, which is challenging or impractical for real-world deployment due to high costs or safety concerns. For example, due to the spatiotemporally unbalanced supply and demand, online reinforcement learning-based order dispatching may significantly impact the revenue of the ride-hailing platform and passenger experience during the policy learning period. Hence, in this work, we develop an offline deep reinforcement learning framework called NondBREM for large-scale order dispatching, which learns policy from only the accumulated logged data to avoid costly and unsafe interactions with the environment. In NondBREM, a Nondeterministic Batch-Constrained Q-learning (NondBCQ) module is developed to reduce the algorithm extrapolation error and a Random Ensemble Mixture (REM) module that integrates multiple value networks with multi-head networks is utilized to improve the model generalization and robustness. Extensive experiments on large-scale real-world ride-hailing datasets show the superiority of our design.

Vehicle Simulation Algorithm for Observations with Variable Dimensions Based on Deep Reinforcement Learning

Vehicle simulation algorithms play a crucial role in enhancing traffic efficiency and safety by predicting and evaluating vehicle behavior in various traffic scenarios. Recently, vehicle simulation algorithms based on reinforcement learning have demonstrated excellent performance in practical tasks due to their ability to exhibit superior performance with zero-shot learning. However, these algorithms face challenges in field adaptation problems when deployed in task sets with variable-dimensional observations, primarily due to the inherent limitations of neural network models. In this paper, we propose a neural network structure accommodating variations in specific dimensions to enhance existing reinforcement learning methods. Building upon this, a scene-compatible vehicle simulation algorithm is designed. We conducted experiments on multiple tasks and scenarios using the Highway-Env traffic environment simulator. The results of our experiments demonstrate that the algorithm can successfully operate on all tasks using a neural network model with fixed shape, even with variable-dimensional observations. Our model exhibits no degradation in simulation performance when compared to the baseline algorithm.

BGRL: Basal Ganglia inspired Reinforcement Learning based framework for deep brain stimulators

Deep Brain Stimulation (DBS) is an implantable medical device used for electrical stimulation to treat neurological disorders. Traditional DBS devices provide fixed frequency pulses, but personalized adjustment of stimulation parameters is crucial for optimal treatment. This paper introduces a Basal Ganglia inspired Reinforcement Learning (BGRL) approach, incorporating a closed-loop feedback mechanism to suppress neural synchrony during neurological fluctuations. The BGRL approach leverages the resemblance between the Basal Ganglia region of brain by incorporating the actor–critic architecture of reinforcement learning (RL). Simulation results demonstrate that BGRL significantly reduces synchronous electrical pulses compared to other standard RL algorithms. BGRL algorithm outperforms existing RL methods in terms of suppression capability and energy consumption, validated through comparisons using ensemble oscillators. Results shown in the paper demonstrate BGRL suppressed the synchronous electrical pulses across three signaling regimes namely regular, chaotic and bursting by 40%, 146% and 40% respectively as compared to soft actor–critic model. BGRL shows promise in effectively suppressing neural synchrony in DBS therapy, providing an efficient alternative to open-loop methodologies.

Counterfactual-attention multi-agent reinforcement learning for joint condition-based maintenance and production scheduling

Maintenance and production scheduling are interactive activities that should be considered simultaneously to maintain production systems’ reliability and high production delivery rate. This study investigates a problem of joint maintenance and production scheduling in a multi-stage hybrid flow shop experiencing machine deterioration processes. To address the structural dependency among production stages and machines, a decentralized partially observable Markov decision process (Dec-POMDP) is considered. However, the large state and action space pose a challenge for existing reinforcement learning methods to provide satisfactory solutions. To overcome this problem, the study proposes a counterfactual attention multi-agent reinforcement learning (CAMARL) solution framework that comprises three functional modules: The attention mechanism module, which adaptively shrinks the dimension of the state space; the action abstraction representation module that deals with the high-dimension of the action space; and the coordination control unit, which accelerates the exploration for the optimal production policy. Numerical experiments demonstrate the effectiveness of the proposed method by comparing it with seven benchmarks under different production scenarios.

Uncertainty-Aware Decision-Making for Autonomous Driving at Uncontrolled Intersections

Reinforcement learning (RL) has been widely used in the decision-making of autonomous vehicles (AVs) in recent studies. However, existing RL methods generally find the optimal policy by maximizing the expectation of future returns, which lacks distributional treatments of risky situations. Additionally, various uncertainties arising from the environment could also cause unreliable decisions, particularly in some complex urban environments. In this paper, the fully parameterized quantile network (FPQN) is utilized to estimate the full return distribution. Then, the conditional value-at-risk (CVaR) is utilized with the return distribution information to generate uncertainty-aware driving behavior. Additionally, an uncontrolled four-way intersection is developed by the Simulation of Urban Mobility (SUMO) simulation platform, which considers both the surrounding vehicles (SVs) and pedestrians. More specifically, to simulate the real-world traffic environment, the uncertainty arising from the occlusion, and the behavior uncertainty of surrounding traffic participants are also considered. The experiment results suggest that the proposed method outperforms the baseline methods in terms of safety. Furthermore, the results also indicate that the proposed method can make reasonable decisions in some challenging driving cases in the presence of uncertainty.

Weighted Policy Constraints for Offline Reinforcement Learning

Offline reinforcement learning (RL) aims to learn policy from the passively collected offline dataset. Applying existing RL methods on the static dataset straightforwardly will raise distribution shift, causing these unconstrained RL methods to fail. To cope with the distribution shift problem, a common practice in offline RL is to constrain the policy explicitly or implicitly close to behavioral policy. However, the available dataset usually contains sub-optimal or inferior actions, constraining the policy near all these actions will make the policy inevitably learn inferior behaviors, limiting the performance of the algorithm. Based on this observation, we propose a weighted policy constraints (wPC) method that only constrains the learned policy to desirable behaviors, making room for policy improvement on other parts. Our algorithm outperforms existing state-of-the-art offline RL algorithms on the D4RL offline gym datasets. Moreover, the proposed algorithm is simple to implement with few hyper-parameters, making the proposed wPC algorithm a robust offline RL method with low computational complexity.

SafeLight: A Reinforcement Learning Method toward Collision-Free Traffic Signal Control

Traffic signal control is safety-critical for our daily life. Roughly one-quarter of road accidents in the U.S. happen at intersections due to problematic signal timing, urging the development of safety-oriented intersection control. However, existing studies on adaptive traffic signal control using reinforcement learning technologies have focused mainly on minimizing traffic delay but neglecting the potential exposure to unsafe conditions. We, for the first time, incorporate road safety standards as enforcement to ensure the safety of existing reinforcement learning methods, aiming toward operating intersections with zero collisions. We have proposed a safety-enhanced residual reinforcement learning method (SafeLight) and employed multiple optimization techniques, such as multi-objective loss function and reward shaping for better knowledge integration. Extensive experiments are conducted using both synthetic and real-world benchmark datasets. Results show that our method can significantly reduce collisions while increasing traffic mobility.

Peano: learning formal mathematical reasoning.

General mathematical reasoning is computationally undecidable, but humans routinely solve new problems. Moreover, discoveries developed over centuries are taught to subsequent generations quickly. What structure enables this, and how might that inform automated mathematical reasoning? We posit that central to both puzzles is the structure of procedural abstractions underlying mathematics. We explore this idea in a case study on five sections of beginning algebra on the Khan Academy platform. To define a computational foundation, we introduce Peano, a theorem-proving environment where the set of valid actions at any point is finite. We use Peano to formalize introductory algebra problems and axioms, obtaining well-defined search problems. We observe existing reinforcement learning methods for symbolic reasoning to be insufficient to solve harder problems. Adding the ability to induce reusable abstractions ('tactics') from its own solutions allows an agent to make steady progress, solving all problems. Furthermore, these abstractions induce an order to the problems, seen at random during training. The recovered order has significant agreement with the expert-designed Khan Academy curriculum, and second-generation agents trained on the recovered curriculum learn significantly faster. These results illustrate the synergistic role of abstractions and curricula in the cultural transmission of mathematics. This article is part of a discussion meeting issue 'Cognitive artificial intelligence'.

A Survey on Reinforcement Learning Methods in Bionic Underwater Robots

Bionic robots possess inherent advantages for underwater operations, and research on motion control and intelligent decision making has expanded their application scope. In recent years, the application of reinforcement learning algorithms in the field of bionic underwater robots has gained considerable attention, and continues to grow. In this paper, we present a comprehensive survey of the accomplishments of reinforcement learning algorithms in the field of bionic underwater robots. Firstly, we classify existing reinforcement learning methods and introduce control tasks and decision making tasks based on the composition of bionic underwater robots. We further discuss the advantages and challenges of reinforcement learning for bionic robots in underwater environments. Secondly, we review the establishment of existing reinforcement learning algorithms for bionic underwater robots from different task perspectives. Thirdly, we explore the existing training and deployment solutions of reinforcement learning algorithms for bionic underwater robots, focusing on the challenges posed by complex underwater environments and underactuated bionic robots. Finally, the limitations and future development directions of reinforcement learning in the field of bionic underwater robots are discussed. This survey provides a foundation for exploring reinforcement learning control and decision making methods for bionic underwater robots, and provides insights for future research.

Learning Diverse Policies with Soft Self-Generated Guidance

Reinforcement learning (RL) with sparse and deceptive rewards is a significant challenge because nonzero rewards are rarely obtained, and hence, the gradient calculated by the agent can be stochastic and without valid information. Recent work demonstrates that using memory buffers of previous experiences can lead to a more efficient learning process. However, existing methods usually require these experiences to be successful and may overly exploit them, which can cause the agent to adopt suboptimal behaviors. This study develops an approach that exploits diverse past trajectories for faster and more efficient online RL, even if these trajectories are suboptimal or not highly rewarded. The proposed algorithm merges a policy improvement step with an additional policy exploration step by using offline demonstration data. The main contribution of this study is that by regarding diverse past trajectories as guidance, instead of imitating them, our method directs its policy to follow and expand past trajectories, while still being able to learn without rewards and gradually approach optimality. Furthermore, a novel diversity measurement is introduced to maintain the diversity of the team and regulate exploration. The proposed algorithm is evaluated on a series of discrete and continuous control tasks with sparse and deceptive rewards. In comparison with the existing RL methods, the experimental results indicate that our proposed algorithm is significantly better than the baseline methods in terms of diverse exploration and avoiding local optima.

Reinforcement Learning for Clinical Applications.

Reinforcement Learning for Clinical Applications.

Architecture for Hierarchical System with Each Learning Structure

In this paper, we propose an architecture for realizing distributed reinforcement learning of distributed controllers for a class of unknown hierarchical systems, where homogeneous subsystems are interconnected through a complete graph. All these controllers consist of two sub-controllers for average and difference dynamics of the system, respectively. First, we show that optimal sub-controllers can be trained individually by a reinforcement learning (RL) method for average/difference data. Due to the smaller-scale of the data, the learning time of the proposed method can be drastically reduced compared to existing RL methods. However, the computation for obtaining the average data requires all-to-all communication among subsystems, which is undesirable in terms of communication costs and security. Hence, by exploiting a distributed consensus observer, we propose an architecture that enables us to learn distributed optimal controllers in a distributed manner. The control performance of the trained controller is shown to be ideally optimal. Moreover, the proposed architecture is completely scalable, i.e., its computational cost is independent from the number of subsystems. The effectiveness is shown through numerical simulations.

Self-Supervised Reinforcement Learning with dual-reward for knowledge-aware recommendation

To improve the recommendation accuracy and offer explanations for recommendations, Reinforcement Learning (RL) has been applied to path reasoning over knowledge graphs. However, in recommendation tasks, most existing RL methods learn the path-finding policy using only a short-term or single reward, leading to a local optimum and losing some potential paths. To address these issues, we propose a Self-Supervised Reinforcement Learning (SSRL) framework combined with dual-reward for knowledge-aware recommendation reasoning over knowledge graphs. Then, we improve Actor–Critic algorithm by using a dual-reward driven strategy, which combines short-term reward with long-term incremental evaluation. The improved algorithm helps the policy guide path reasoning in an overall situation. In addition, to find the most potential paths, in the improved Actor–Critic algorithm, a loss constraint of each sample is used as a reinforced signal to update the gradients. With some improvements against baselines, experimental results demonstrate the effectiveness of our framework.

Optimal stroke learning with policy gradient approach for robotic table tennis

Learning to play table tennis is a challenging task for robots, as a wide variety of strokes required. Recent advances have shown that deep Reinforcement Learning (RL) is able to successfully learn the optimal actions in a simulated environment. However, the applicability of RL in real scenarios remains limited due to the high exploration effort. In this work, we propose a realistic simulation environment in which multiple models are built for the dynamics of the ball and the kinematics of the robot. Instead of training an end-to-end RL model, a novel policy gradient approach with TD3 backbone is proposed to learn the racket strokes based on the predicted state of the ball at the hitting time. In the experiments, we show that the proposed approach significantly outperforms the existing RL methods in simulation. Furthermore, to cross the domain from simulation to reality, we adopt an efficient retraining method and test it in three real scenarios. The resulting success rate is 98% and the distance error is around 24.9 cm. The total training time is about 1.5 hours.

Automatic index selection with learned cost estimator

Choosing useful indexes for a relational database is important for efficient query optimization. However, current methods tend to use the cost estimated by the optimizer in the database management system (DBMS) to measure the benefits of the index and, due to inaccurate cost estimates, do not bring to the most optimal solution. In addition, existing reinforcement learning methods treat the creation of different indexes as independent actions, ignoring the relationships among indexes, which may bring to unnecessary training costs. To address these problems, we propose DeepIndex, an automatic index selector with a learning-based cost estimator to improve the quality of index selection. To accurately estimate the benefit of an index, we design a learning-based cost estimator to predict the execution time of queries on certain indexes. In particular, we treat query plans as graphs and develop a model based on graph convolutional network (GCN) to learn features from queries and indexes. After that, we design a reinforcement-learning-based index selection model considering the relationships among indexes and combine our cost estimator to select indexes. Extensive experiments on two benchmark workloads JOB and TPC-DS show that our model performs better than state-of-the-art models with the lowest storage costs, and our lowest relative execution cost outperforms the baselines by 38.59% on JOB and 10.42% on TPC-DS.