Search

This paper presents the design of an optimal controller for solving tracking problems subject to unmeasurable disturbances and unknown system dynamics using reinforcement learning (RL). Many existing RL control methods take disturbance into account by directly measuring it and manipulating it for exploration during the learning process, thereby preventing any disturbance induced bias in the control estimates. However, in most practical scenarios, disturbance is neither measurable nor manipulable. The main contribution of this article is the introduction of a combination of a bias compensation mechanism and the integral action in the Q-learning framework to remove the need to measure or manipulate the disturbance, while preventing disturbance induced bias in the optimal control estimates. A bias compensated Q-learning scheme is presented that learns the disturbance induced bias terms separately from the optimal control parameters and ensures the convergence of the control parameters to the optimal solution even in the presence of unmeasurable disturbances. Both state feedback and output feedback algorithms are developed based on policy iteration (PI) and value iteration (VI) that guarantee the convergence of the tracking error to zero. The feasibility of the design is validated on a practical optimal control application of a heating, ventilating, and air conditioning (HVAC) zone controller. [ABSTRACT FROM AUTHOR]

In this paper, an adaptive admittance control scheme is developed for robots to interact with time-varying environments. Admittance control is adopted to achieve a compliant physical robot–environment interaction, and the uncertain environment with time-varying dynamics is defined as a linear system. A critic learning method is used to obtain the desired admittance parameters based on the cost function composed of interaction force and trajectory tracking without the knowledge of the environmental dynamics. To deal with dynamic uncertainties in the control system, a neural-network (NN)-based adaptive controller with a dynamic learning framework is developed to guarantee the trajectory tracking performance. Experiments are conducted and the results have verified the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

The learning automaton (LA) that simulates the interaction between an intelligent agent and a stochastic environment to learn the optimal action is an important tool in reinforcement learning. Being confronted with an unknown environment, most learning automata have more than one parameters to be tuned during a pretraining process in which the LA interacts with the environment. Only after the parameters are tuned properly, an LA can act most properly during the training procedure to obtain the optimal behavior. The cost of parameter tuning can be enormous, e.g., possibly millions of interactions are required to seek the best parameter configuration. Therefore, the parameter-free LA that uses identical parameters for every environment and saves further tuning has become the hot spot of this research. This article proposes an efficient parameter-free learning automaton (EPFLA) that depends on a separating function (SF). Taking advantage of both frequentist inference and Bayesian inference, the SF plays a dual role in the proposed scheme: 1) evaluating the difference in performance between actions in the environment and 2) exploring actions by coining an action selection strategy. A proof is provided to ensure the $\epsilon $ -optimality of EPFLA. Comprehensive comparisons verify the privileges of EPFLA over both parameter-based schemes and existing parameter-free schemes. [ABSTRACT FROM AUTHOR]

Reinforcement learning algorithms, such as hindsight experience replay (HER) and hindsight goal generation (HGG), have been able to solve challenging robotic manipulation tasks in multigoal settings with sparse rewards. HER achieves its training success through hindsight replays of past experience with heuristic goals but underperforms in challenging tasks in which goals are difficult to explore. HGG enhances HER by selecting intermediate goals that are easy to achieve in the short term and promising to lead to target goals in the long term. This guided exploration makes HGG applicable to tasks in which target goals are far away from the object’s initial position. However, the vanilla HGG is not applicable to manipulation tasks with obstacles because the Euclidean metric used for HGG is not an accurate distance metric in such an environment. Although, with the guidance of a handcrafted distance grid, grid-based HGG can solve manipulation tasks with obstacles, a more feasible method that can solve such tasks automatically is still in demand. In this article, we propose graph-based hindsight goal generation (G-HGG), an extension of HGG selecting hindsight goals based on shortest distances in an obstacle-avoiding graph, which is a discrete representation of the environment. We evaluated G-HGG on four challenging manipulation tasks with obstacles, where significant enhancements in both sample efficiency and overall success rate are shown over HGG and HER. Videos can be viewed at https://videoviewsite.wixsite.com/ghgg. [ABSTRACT FROM AUTHOR]

Fusing low dynamic range (LDR) for high dynamic range (HDR) images has gained a lot of attention, especially to achieve real-world application significance when the hardware resources are limited to capture images with different exposure times. However, existing HDR image generation by picking the best parts from each LDR image often yields unsatisfactory results due to either the lack of input images or well-exposed contents. To overcome this limitation, we model the HDR image generation process in two-exposure fusion as a deep reinforcement learning problem and learn an online compensating representation to fuse with LDR inputs for HDR image generation. Moreover, we build a two-exposure dataset with reference HDR images from a public multiexposure dataset that has not yet been normalized to train and evaluate the proposed model. By assessing the built dataset, we show that our reinforcement HDR image generation significantly outperforms other competing methods under different challenging scenarios, even with limited well-exposed contents. More experimental results on a no-reference multiexposure image dataset demonstrate the generality and effectiveness of the proposed model. To the best of our knowledge, this is the first work to use a reinforcement-learning-based framework for an online compensating representation in two-exposure image fusion. [ABSTRACT FROM AUTHOR]

How to generate more revenues is crucial to cloud providers. Evidences from the Amazon cloud system indicate that “dynamic pricing” would be more profitable than “static pricing.” The challenges are: How to set the price in real-time so to maximize revenues? How to estimate the price dependent demand so to optimize the pricing decision? We first design a discrete-time based dynamic pricing scheme and formulate a Markov decision process to characterize the evolving dynamics of the price-dependent demand. We formulate a revenue maximization framework to determine the optimal price and theoretically characterize the “structure” of the optimal revenue and optimal price. We apply the $Q$ -learning to infer the optimal price from historical transaction data and derive sufficient conditions on the model to guarantee its convergence to the optimal price, but it converges slowly. To speed up the convergence, we incorporate the structure of the optimal revenue that we obtained earlier, leading to the VpQ-learning ($Q$ -learning with value projection) algorithm. We derive sufficient conditions, under which the VpQ-learning algorithm converges to the optimal policy. Experiments on a real-world dataset show that the VpQ-learning algorithm outperforms a variety of baselines, i.e., improves the revenue by as high as 50% over the $Q$ -learning, speedy $Q$ -learning, and adaptive real-time dynamic programming (ARTDP), and by as high as 20% over the fixed pricing scheme. [ABSTRACT FROM AUTHOR]

In reinforcement learning (RL), function approximation errors are known to easily lead to the $Q$ -value overestimations, thus greatly reducing policy performance. This article presents a distributional soft actor–critic (DSAC) algorithm, which is an off-policy RL method for continuous control setting, to improve the policy performance by mitigating $Q$ -value overestimations. We first discover in theory that learning a distribution function of state–action returns can effectively mitigate $Q$ -value overestimations because it is capable of adaptively adjusting the update step size of the $Q$ -value function. Then, a distributional soft policy iteration (DSPI) framework is developed by embedding the return distribution function into maximum entropy RL. Finally, we present a deep off-policy actor–critic variant of DSPI, called DSAC, which directly learns a continuous return distribution by keeping the variance of the state–action returns within a reasonable range to address exploding and vanishing gradient problems. We evaluate DSAC on the suite of MuJoCo continuous control tasks, achieving the state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Reinforcement learning (RL) is a promising technique for designing a model-free controller by interacting with the environment. Several researchers have applied RL to autonomous underwater vehicles (AUVs) for motion control, such as trajectory tracking. However, the existing RL-based controller usually assumes that the unknown AUV dynamics keep invariant during the operation period, limiting its further application in the complex underwater environment. In this article, a novel meta-RL-based control scheme is proposed for trajectory tracking control of AUV in the presence of unknown and time-varying dynamics. To this end, we divide the tracking task for AUV with time-varying dynamics into multiple specific tasks with fixed time-varying dynamics, to which we apply meta-RL for training to distill the general control policy. The obtained control policy can transfer to the testing phase with high adaptability. Inspired by the line-of-sight (LOS) tracking rule, we formulate each specific task as a Markov decision process (MDP) with a well-designed state and reward function. Furthermore, a novel policy network with an attention module is proposed to extract the hidden information of AUV dynamics. The simulation environment with time-varying dynamics is established, and the simulation results reveal the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

This article presents an iterative data-driven algorithm for solving dynamic multiobjective (MO) optimal control problems arising in control of nonlinear continuous-time systems. It is first shown that the Hamiltonian functional corresponding to each objective can be leveraged to compare the performance of admissible policies. Hamiltonian inequalities are then used for which their satisfaction guarantees satisfying the objectives’ aspirations. Relaxed Hamilton–Jacobi–Bellman (HJB) equations in terms of HJB inequalities are then solved in a dynamic constrained MO framework to find Pareto optimal solutions. Relation to satisficing (good enough) decision-making framework is shown. A sum-of-square (SOS)-based iterative algorithm is developed to solve the formulated aspiration-satisfying MO optimization. To obviate the requirement of complete knowledge of the system dynamics, a data-driven satisficing reinforcement learning approach is proposed to solve the SOS optimization problem in real time using only the information of the system trajectories measured during a time interval without having full knowledge of the system dynamics. Finally, two simulation examples are utilized to verify the analytical results of the proposed algorithm. [ABSTRACT FROM AUTHOR]

We are motivated by the real challenges presented in a human–robot system to develop new designs that are efficient at data level and with performance guarantees, such as stability and optimality at system level. Existing approximate/adaptive dynamic programming (ADP) results that consider system performance theoretically are not readily providing practically useful learning control algorithms for this problem, and reinforcement learning (RL) algorithms that address the issue of data efficiency usually do not have performance guarantees for the controlled system. This study fills these important voids by introducing innovative features to the policy iteration algorithm. We introduce flexible policy iteration (FPI), which can flexibly and organically integrate experience replay and supplemental values from prior experience into the RL controller. We show system-level performances, including convergence of the approximate value function, (sub)optimality of the solution, and stability of the system. We demonstrate the effectiveness of the FPI via realistic simulations of the human–robot system. It is noted that the problem we face in this study may be difficult to address by design methods based on classical control theory as it is nearly impossible to obtain a customized mathematical model of a human–robot system either online or offline. The results we have obtained also indicate the great potential of RL control to solving realistic and challenging problems with high-dimensional control inputs. [ABSTRACT FROM AUTHOR]

In this article, we study the problem of guaranteed display ads (GDAs) allocation, which requires proactively allocate display ads to different impressions to fulfill their impression demands indicated in the contracts. Existing methods for this problem either assume the impressions that are static or solely consider a specific ad’s benefits. Thus, it is hard to generalize to the industrial production scenario where the impressions are dynamical and large-scale, and the overall allocation optimality of all the considered GDAs is required. To bridge this gap, we formulate this problem as a sequential decision-making problem in the scope of multiagent reinforcement learning (MARL), by assigning an allocation agent to each ad and coordinating all the agents for allocating GDAs. The inputs are the states (e.g., the demands of the ad and the remaining time steps for displaying the ads) of each ad and the impressions at different time steps, and the outputs are the display ratios of each ad for each impression. Specifically, we propose a novel hierarchical MARL (HMARL) method that creates hierarchies over the agent policies to handle a large number of ads and the dynamics of impressions. HMARL contains: 1) a manager policy to navigate the agent to choose an appropriate subpolicy and 2) a set of subpolicies that let the agents perform diverse conditioning on their states. Extensive experiments on three real-world data sets from the Tencent advertising platform with tens of millions of records demonstrate significant improvements of HMARL over state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Direct policy search (DPS) is emerging as one of the most effective and widely applied reinforcement learning (RL) methods to design optimal control policies for multiobjective Markov decision processes (MOMDPs). Traditionally, DPS defines the control policy within a preselected functional class and searches its optimal parameterization with respect to a given set of objectives. The functional class should be tailored to the problem at hand and its selection is crucial, as it determines the search space within which solutions can be found. In MOMDPs problems, a different objective tradeoff determines a different fitness landscape, requiring a tradeoff-dynamic functional class selection. Yet, in state-of-the-art applications, the policy class is generally selected a priori and kept constant across the multidimensional objective space. In this work, we present a novel policy search routine called neuro-evolutionary multiobjective DPS (NEMODPS), which extends the DPS problem formulation to conjunctively search the policy functional class and its parameterization in a hyperspace containing policy architectures and coefficients. NEMODPS begins with a population of minimally structured approximating networks and progressively builds more sophisticated architectures by topological and parametrical mutation and crossover, and selection of the fittest individuals concerning multiple objectives. We tested NEMODPS for the problem of designing the control policy of a multipurpose water system. Numerical results show that the tradeoff-dynamic structural and parametrical policy search of NEMODPS is consistent across multiple runs, and outperforms the solutions designed via traditional DPS with predefined policy topologies. [ABSTRACT FROM AUTHOR]

In this article, we investigate the routing problem of packet networks through multiagent reinforcement learning (RL), which is a very challenging topic in distributed and autonomous networked systems. In specific, the routing problem is modeled as a networked multiagent partially observable Markov decision process (MDP). Since the MDP of a network node is not only affected by its neighboring nodes’ policies but also the network traffic demand, it becomes a multitask learning problem. Inspired by recent success of RL and metalearning, we propose two novel model-free multiagent RL algorithms, named multiagent proximal policy optimization (MAPPO) and multiagent metaproximal policy optimization (meta-MAPPO), to optimize the network performances under fixed and time-varying traffic demand, respectively. A practicable distributed implementation framework is designed based on the separability of exploration and exploitation in training MAPPO. Compared with the existing routing optimization policies, our simulation results demonstrate the excellent performances of the proposed algorithms. [ABSTRACT FROM AUTHOR]

We study the decentralized event-driven control problem of nonlinear dynamical systems with mismatched interconnections and asymmetric input constraints. To begin with, by introducing a discounted cost function for each auxiliary subsystem, we transform the decentralized event-driven constrained control problem into a group of nonlinear $H_{2}$ -constrained optimal control problems. Then, we develop the event-driven Hamilton–Jacobi–Bellman equations (ED-HJBEs), which arise in the nonlinear $H_{2}$ -constrained optimal control problems. Meanwhile, we demonstrate that all the solutions of the ED-HJBEs together keep the overall system stable in the sense of uniform ultimate boundedness (UUB). To solve the ED-HJBEs, we build a critic-only architecture under the framework of adaptive critic designs. The architecture only employs critic neural networks and updates their weight vectors via the gradient descent method. After that, based on the Lyapunov approach, we prove that the UUB stability of all signals in the closed-loop auxiliary subsystems is assured. Finally, simulations of an illustrated nonlinear interconnected plant are provided to validate the present designs. [ABSTRACT FROM AUTHOR]

In this article, a data-driven distributed control method is proposed to solve the cooperative optimal output regulation problem of leader–follower multiagent systems. Different from traditional studies on cooperative output regulation, a distributed adaptive internal model is originally developed, which includes a distributed internal model and a distributed observer to estimate the leader’s dynamics. Without relying on the dynamics of multiagent systems, we have proposed two reinforcement learning algorithms, policy iteration and value iteration, to learn the optimal controller through online input and state data, and estimated values of the leader’s state. By combining these methods, we have established a basis for connecting data-distributed control methods with adaptive dynamic programming approaches in general since these are the theoretical foundation from which they are built. [ABSTRACT FROM AUTHOR]

Recent studies in using deep learning (DL) to solve routing problems focus on construction heuristics, whose solutions are still far from optimality. Improvement heuristics have great potential to narrow this gap by iteratively refining a solution. However, classic improvement heuristics are all guided by handcrafted rules that may limit their performance. In this article, we propose a deep reinforcement learning framework to learn the improvement heuristics for routing problems. We design a self-attention-based deep architecture as the policy network to guide the selection of the next solution. We apply our method to two important routing problems, i.e., the traveling salesman problem (TSP) and the capacitated vehicle routing problem (CVRP). Experiments show that our method outperforms state-of-the-art DL-based approaches. The learned policies are more effective than the traditional handcrafted ones and can be further enhanced by simple diversifying strategies. Moreover, the policies generalize well to different problem sizes, initial solutions, and even real-world data set. [ABSTRACT FROM AUTHOR]

Balancing the supply and demand for ride-sourcing companies is a challenging issue, especially with real-time requests and stochastic traffic conditions of large-scale congested road networks. To tackle this challenge, this article proposes a robust and scalable approach that integrates reinforcement learning (RL) and a centralized programming (CP) structure to promote real-time taxi operations. Both real-time order matching decisions and vehicle relocation decisions at the microscopic network scale are integrated within a Markov decision process framework. The RL component learns the decomposed state-value function, which represents the taxi drivers’ experience, the off-line historical demand pattern, and the traffic network congestion. The CP component plans nonmyopic decisions for drivers collectively under the prescribed system constraints to explicitly realize cooperation. Furthermore, to circumvent sparse reward and sample imbalance problems over the microscopic road network, this article proposed a temporal-difference learning algorithm with prioritized gradient descent and adaptive exploration techniques. A simulator is built and trained with the Manhattan road network and New York City yellow taxi data to simulate the real-time vehicle dispatching environment. Both centralized and decentralized taxi dispatching policies are examined with the simulator. This case study shows that the proposed approach can further improve taxi drivers’ profits while reducing customers’ waiting times compared to several existing vehicle dispatching algorithms. [ABSTRACT FROM AUTHOR]

Perturbation has a positive effect, as it contributes to the stability of neural systems through adaptation and robustness. For example, deep reinforcement learning generally engages in exploratory behavior by injecting noise into the action space and network parameters. It can consistently increase the agent’s exploration ability and lead to richer sets of behaviors. Evolutionary strategies also apply parameter perturbations, which makes network architecture robust and diverse. Our main concern is whether the notion of synaptic perturbation introduced in a spiking neural network (SNN) is biologically relevant or if novel frameworks and components are desired to account for the perturbation properties of artificial neural systems. In this work, we first review part of the locality-sensitive hashing (LSH) of similarity search, the FLY algorithm, as recently published in Science, and propose an improved architecture, time-shifted spiking LSH (TS-SLSH), with the consideration of temporal perturbations of the firing moments of spike pulses. Experiment results show promising performance of the proposed method and demonstrate its generality to various spiking neuron models. Therefore, we expect temporal perturbation to play an active role in SNN performance. [ABSTRACT FROM AUTHOR]

Deep neural network-based systems are now state-of-the-art in many robotics tasks, but their application in safety-critical domains remains dangerous without formal guarantees on network robustness. Small perturbations to sensor inputs (from noise or adversarial examples) are often enough to change network-based decisions, which was recently shown to cause an autonomous vehicle to swerve into another lane. In light of these dangers, numerous algorithms have been developed as defensive mechanisms from these adversarial inputs, some of which provide formal robustness guarantees or certificates. This work leverages research on certified adversarial robustness to develop an online certifiably robust for deep reinforcement learning algorithms. The proposed defense computes guaranteed lower bounds on state-action values during execution to identify and choose a robust action under a worst case deviation in input space due to possible adversaries or noise. Moreover, the resulting policy comes with a certificate of solution quality, even though the true state and optimal action are unknown to the certifier due to the perturbations. The approach is demonstrated on a deep Q-network (DQN) policy and is shown to increase robustness to noise and adversaries in pedestrian collision avoidance scenarios, a classic control task, and Atari Pong. This article extends our prior work with new performance guarantees, extensions to other reinforcement learning algorithms, expanded results aggregated across more scenarios, an extension into scenarios with adversarial behavior, comparisons with a more computationally expensive method, and visualizations that provide intuition about the robustness algorithm. [ABSTRACT FROM AUTHOR]

With the rapid evolution of wireless mobile devices, there emerges an increased need to design effective collaboration mechanisms between intelligent agents to gradually approach the final collective objective by continuously learning from the environment based on their individual observations. In this regard, independent reinforcement learning (IRL) is often deployed in multiagent collaboration to alleviate the problem of a nonstationary learning environment. However, behavioral strategies of intelligent agents in IRL can be formulated only upon their local individual observations of the global environment, and appropriate communication mechanisms must be introduced to reduce their behavioral localities. In this article, we address the problem of communication between intelligent agents in IRL by jointly adopting mechanisms with two different scales. For the large scale, we introduce the stigmergy mechanism as an indirect communication bridge between independent learning agents, and carefully design a mathematical method to indicate the impact of digital pheromone. For the small scale, we propose a conflict-avoidance mechanism between adjacent agents by implementing an additionally embedded neural network to provide more opportunities for participants with higher action priorities. In addition, we present a federal training method to effectively optimize the neural network of each agent in a decentralized manner. Finally, we establish a simulation scenario in which a number of mobile agents in a certain area move automatically to form a specified target shape. Extensive simulations demonstrate the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

A central capability of a long-lived reinforcement learning (RL) agent is to incrementally adapt its behavior as its environment changes and to incrementally build upon previous experiences to facilitate future learning in real-world scenarios. In this article, we propose lifelong incremental reinforcement learning (LLIRL), a new incremental algorithm for efficient lifelong adaptation to dynamic environments. We develop and maintain a library that contains an infinite mixture of parameterized environment models, which is equivalent to clustering environment parameters in a latent space. The prior distribution over the mixture is formulated as a Chinese restaurant process (CRP), which incrementally instantiates new environment models without any external information to signal environmental changes in advance. During lifelong learning, we employ the expectation–maximization (EM) algorithm with online Bayesian inference to update the mixture in a fully incremental manner. In EM, the E-step involves estimating the posterior expectation of environment-to-cluster assignments, whereas the M-step updates the environment parameters for future learning. This method allows for all environment models to be adapted as necessary, with new models instantiated for environmental changes and old models retrieved when previously seen environments are encountered again. Simulation experiments demonstrate that LLIRL outperforms relevant existing methods and enables effective incremental adaptation to various dynamic environments for lifelong learning. [ABSTRACT FROM AUTHOR]

Attention has been shown highly effective for modeling sequences, capturing the more informative parts in learning a deep representation. However, recent studies show that the attention values do not always coincide with intuition in tasks, such as machine translation and sentiment classification. In this study, we consider using deep reinforcement learning to automatically optimize attention distribution during the minimization of end task training losses. With more sufficient environment states, iterative actions are taken to adjust attention weights so that more informative words receive more attention automatically. Results on different tasks and different attention networks demonstrate that our model is of great effectiveness in improving the end task performances, yielding more reasonable attention distribution. The more in-depth analysis further reveals that our retrofitting method can help to bring explainability for baseline attention. [ABSTRACT FROM AUTHOR]

In this work, time-driven learning refers to the machine learning method that updates parameters in a prediction model continuously as new data arrives. Among existing approximate dynamic programming (ADP) and reinforcement learning (RL) algorithms, the direct heuristic dynamic programming (dHDP) has been shown an effective tool as demonstrated in solving several complex learning control problems. It continuously updates the control policy and the critic as system states continuously evolve. It is therefore desirable to prevent the time-driven dHDP from updating due to insignificant system event such as noise. Toward this goal, we propose a new event-driven dHDP. By constructing a Lyapunov function candidate, we prove the uniformly ultimately boundedness (UUB) of the system states and the weights in the critic and the control policy networks. Consequently, we show the approximate control and cost-to-go function approaching Bellman optimality within a finite bound. We also illustrate how the event-driven dHDP algorithm works in comparison to the original time-driven dHDP. [ABSTRACT FROM AUTHOR]

Data in many practical problems are acquired according to decisions or actions made by users or experts to achieve specific goals. For instance, policies in the mind of biologists during the intervention process in genomics and metagenomics are often reflected in available data in these domains, or data in cyber–physical systems are often acquired according to actions/decisions made by experts/engineers for purposes, such as control or stabilization. Quantification of experts’ policies through available data, which is also known as reward function learning, has been discussed extensively in the literature in the context of inverse reinforcement learning (IRL). However, most of the available techniques come short to deal with practical problems due to the following main reasons: 1) lack of scalability: arising from incapability or poor performance of existing techniques in dealing with large systems and 2) lack of reliability: coming from the incapability of the existing techniques to properly learn the optimal reward function during the learning process. Toward this, in this brief, we propose a multifidelity Bayesian optimization (MFBO) framework that significantly scales the learning process of a wide range of existing IRL techniques. The proposed framework enables the incorporation of multiple approximators and efficiently takes their uncertainty and computational costs into account to balance exploration and exploitation during the learning process. The proposed framework’s high performance is demonstrated through genomics, metagenomics, and sets of random simulated problems. [ABSTRACT FROM AUTHOR]

In this article, we consider a subclass of partially observable Markov decision process (POMDP) problems which we termed confounding POMDPs. In these types of POMDPs, temporal difference (TD)-based reinforcement learning (RL) algorithms struggle, as TD error cannot be easily derived from observations. We solve these types of problems using a new bio-inspired neural architecture that combines a modulated Hebbian network (MOHN) with deep Q-network (DQN), which we call modulated Hebbian plus Q-network architecture (MOHQA). The key idea is to use a Hebbian network with rarely correlated bio-inspired neural traces to bridge temporal delays between actions and rewards when confounding observations and sparse rewards result in inaccurate TD errors. In MOHQA, DQN learns low-level features and control, while the MOHN contributes to high-level decisions by associating rewards with past states and actions. Thus, the proposed architecture combines two modules with significantly different learning algorithms, a Hebbian associative network and a classical DQN pipeline, exploiting the advantages of both. Simulations on a set of POMDPs and on the Malmo environment show that the proposed algorithm improved DQN’s results and even outperformed control tests with advantage-actor critic (A2C), quantile regression DQN with long short-term memory (QRDQN + LSTM), Monte Carlo policy gradient (REINFORCE), and aggregated memory for reinforcement learning (AMRL) algorithms on most difficult POMDPs with confounding stimuli and sparse rewards. [ABSTRACT FROM AUTHOR]

In this article, a novel integral reinforcement learning (IRL) algorithm is proposed to solve the optimal control problem for continuous-time nonlinear systems with unknown dynamics. The main challenging issue in learning is how to reject the oscillation caused by the externally added probing noise. This article challenges the issue by embedding an auxiliary trajectory that is designed as an exciting signal to learn the optimal solution. First, the auxiliary trajectory is used to decompose the state trajectory of the controlled system. Then, by using the decoupled trajectories, a model-free policy iteration (PI) algorithm is developed, where the policy evaluation step and the policy improvement step are alternated until convergence to the optimal solution. It is noted that an appropriate external input is introduced at the policy improvement step to eliminate the requirement of the input-to-state dynamics. Finally, the algorithm is implemented on the actor–critic structure. The output weights of the critic neural network (NN) and the actor NN are updated sequentially by the least-squares methods. The convergence of the algorithm and the stability of the closed-loop system are guaranteed. Two examples are given to show the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

The Nash equilibrium is an important concept in game theory. It describes the least exploitability of one player from any opponents. We combine game theory, dynamic programming, and recent deep reinforcement learning (DRL) techniques to online learn the Nash equilibrium policy for two-player zero-sum Markov games (TZMGs). The problem is first formulated as a Bellman minimax equation, and generalized policy iteration (GPI) provides a double-loop iterative way to find the equilibrium. Then, neural networks are introduced to approximate Q functions for large-scale problems. An online minimax Q network learning algorithm is proposed to train the network with observations. Experience replay, dueling network, and double Q-learning are applied to improve the learning process. The contributions are twofold: 1) DRL techniques are combined with GPI to find the TZMG Nash equilibrium for the first time and 2) the convergence of the online learning algorithm with a lookup table and experience replay is proven, whose proof is not only useful for TZMGs but also instructive for single-agent Markov decision problems. Experiments on different examples validate the effectiveness of the proposed algorithm on TZMG problems. [ABSTRACT FROM AUTHOR]

Deep reinforcement learning (DRL) has been shown to be successful in many application domains. Combining recurrent neural networks (RNNs) and DRL further enables DRL to be applicable in non-Markovian environments by capturing temporal information. However, training of both DRL and RNNs is known to be challenging requiring a large amount of training data to achieve convergence. In many targeted applications, such as those used in the fifth-generation (5G) cellular communication, the environment is highly dynamic, while the available training data is very limited. Therefore, it is extremely important to develop DRL strategies that are capable of capturing the temporal correlation of the dynamic environment requiring limited training overhead. In this article, we introduce the deep echo state Q-network (DEQN) that can adapt to the highly dynamic environment in a short period of time with limited training data. We evaluate the performance of the introduced DEQN method under the dynamic spectrum sharing (DSS) scenario, which is a promising technology in 5G and future 6G networks to increase the spectrum utilization. Compared with conventional spectrum management policy that grants a fixed spectrum band to a single system for exclusive access, DSS allows the secondary system to share the spectrum with the primary system. Our work sheds light on the application of an efficient DRL framework in highly dynamic environments with limited available training data. [ABSTRACT FROM AUTHOR]

This article is concerned with the optimal synchronization problem for discrete-time nonlinear heterogeneous multiagent systems (MASs) with an active leader. To overcome the difficulty in the derivation of the optimal control protocols for these systems, we develop an observer-based adaptive synchronization control approach, including the designs of a distributed observer and a distributed model reference adaptive controller with no prior knowledge of all agents’ dynamics. To begin with, for the purpose of estimating the state of a nonlinear active leader for each follower, an adaptive neural network distributed observer is designed. Such an observer serves as a reference model in the distributed model reference adaptive control (MRAC). Then, a reinforcement learning-based distributed MRAC algorithm is presented to make every follower track its corresponding reference model on behavior in real time. In this algorithm, a distributed actor–critic network is employed to approximate the optimal distributed control protocols and the cost function. Through convergence analysis, the overall observer estimation error, the model reference tracking error, and the weight estimation errors are proved to be uniformly ultimately bounded. The developed approach further achieves the synchronization by means of synthesizing these results. The effectiveness of the developed approach is verified through a numerical example. [ABSTRACT FROM AUTHOR]

Existing deep reinforcement learning (RL) are devoted to research applications on video games, e.g., The Open Racing Car Simulator (TORCS) and Atari games. However, it remains under-explored for vision-based autonomous urban driving navigation (VB-AUDN). VB-AUDN requires a sophisticated agent working safely in structured, changing, and unpredictable environments; otherwise, inappropriate operations may lead to irreversible or catastrophic damages. In this work, we propose a deductive RL (DeRL) to address this challenge. A deduction reasoner (DR) is introduced to endow the agent with ability to foresee the future and to promote policy learning. Specifically, DR first predicts future transitions through a parameterized environment model. Then, DR conducts self-assessment at the predicted trajectory to perceive the consequences of current policy resulting in a more reliable decision-making process. Additionally, a semantic encoder module (SEM) is designed to extract compact driving representation from the raw images, which is robust to the changes of the environment. Extensive experimental results demonstrate that DeRL outperforms the state-of-the-art model-free RL approaches on the public CAR Learning to Act (CARLA) benchmark and presents a superior performance on success rate and driving safety for goal-directed navigation. [ABSTRACT FROM AUTHOR]

Learning to adapt to a series of different goals in visual navigation is challenging. In this work, we present a model-embedded actor–critic architecture for the multigoal visual navigation task. To enhance the task cooperation in multigoal learning, we introduce two new designs to the reinforcement learning scheme: inverse dynamics model (InvDM) and multigoal colearning (MgCl). Specifically, InvDM is proposed to capture the navigation-relevant association between state and goal and provide additional training signals to relieve the sparse reward issue. MgCl aims at improving the sample efficiency and supports the agent to learn from unintentional positive experiences. Besides, to further improve the scene generalization capability of the agent, we present an enhanced navigation model that consists of two self-supervised auxiliary task modules. The first module, which is named path closed-loop detection, helps to understand whether the state has been experienced. The second one, namely the state-target matching module, tries to figure out the difference between state and goal. Extensive results on the interactive platform AI2-THOR demonstrate that the agent trained with the proposed method converges faster than state-of-the-art methods while owning good generalization capability. The video demonstration is available at https://vsislab.github.io/mgvn. [ABSTRACT FROM AUTHOR]

In this article, the model-free robust formation control problem is addressed for cooperative underactuated quadrotors involving unknown nonlinear dynamics and disturbances. Based on the hierarchical control scheme and the reinforcement learning theory, a robust controller is proposed without knowledge of each quadrotor dynamics, consisting of a distributed observer to estimate the position state of the leader, a position controller to achieve the desired formation, and an attitude controller to control the rotational motion. Simulation results on the multiquadrotor system confirm the effectiveness of the proposed model-free robust formation control method. [ABSTRACT FROM AUTHOR]

This article applies a singular perturbation theory to solve an optimal linear quadratic tracker problem for a continuous-time two-time-scale process. Previously, singular perturbation was applied for system regulation. It is shown that the two-time-scale tracking problem can be separated into a linear–quadratic tracker (LQT) problem for the slow system and a linear–quadratic regulator (LQR) problem for the fast system. We prove that the solutions to these two reduced-order control problems can approximate the LQT solution of the original control problem. The reduced-order slow LQT and fast LQR control problems are solved by off-policy integral reinforcement learning (IRL) using only measured data from the system. To test the effectiveness of the proposed method, we use an industrial thickening process as a simulation example and compare our method to a method with the known system model and a method without time-scale separation. [ABSTRACT FROM AUTHOR]