Recent Progress in Reinforcement Learning and Adaptive Dynamic Programming for Advanced Control Applications

Reinforcement learning(RL) has roots in dynamic programming and it is called adaptive/approximate dynamic programming(ADP) within the control community. This paper reviews recent developments in ADP along with RL and its applications to various advanced control fields. First, the background of the development of ADP is described, emphasizing the significance of regulation and tracking control problems. Some effective offline and online algorithms for ADP/adaptive critic control are displayed, where the main results towards discrete-time systems and continuous-time systems are surveyed, respectively.Then, the research progress on adaptive critic control based on the event-triggered framework and under uncertain environment is discussed, respectively, where event-based design, robust stabilization, and game design are reviewed. Moreover, the extensions of ADP for addressing control problems under complex environment attract enormous attention. The ADP architecture is revisited under the perspective of data-driven and RL frameworks,showing how they promote ADP formulation significantly.Finally, several typical control applications with respect to RL and ADP are summarized, particularly in the fields of wastewater treatment processes and power systems, followed by some general prospects for future research. Overall, the comprehensive survey on ADP and RL for advanced control applications has d emonstrated its remarkable potential within the artificial intelligence era. In addition, it also plays a vital role in promoting environmental protection and industrial intelligence.

Adaptive Optimal Discrete-Time Output-Feedback Using an Internal Model Principle and Adaptive Dynamic Programming

In order to address the output feedback issue for linear discrete-time systems, this work suggests a brand-new adaptive dynamic programming(ADP) technique based on the internal model principle(IMP). The proposed method, termed as IMP-ADP, does not require complete state feedback-merely the measurement of input and output data. More specifically, based on the IMP, the output control problem can first be converted into a stabilization problem. We then design an observer to reproduce the full state of the system by measuring the inputs and outputs. Moreover, this technique includes both a policy iteration algorithm and a value iteration algorithm to determine the optimal feedback gain without using a dynamic system model. It is important that with this concept one does not need to solve the regulator equation. Finally, this control method was tested on an inverter system of grid-connected LCLs to demonstrate that the proposed method provides the desired performance in terms of both tracking and disturbance rejection.

Event-based performance guaranteed tracking control for constrained nonlinear system via adaptive dynamic programming method

An optimal tracking control problem for a class of nonlinear systems with guaranteed performance and asymmetric input constraints is discussed in this paper.The control policy is implemented by adaptive dynamic programming(ADP)algorithm under two event-based triggering mechanisms.It is often challenging to design an optimal control law due to the system deviation caused by asymmetric input constraints.First,a prescribed performance control technique is employed to guarantee the tracking errors within predetermined boundaries.Subsequently,considering the asymmetric input constraints,a discounted non-quadratic cost function is introduced.Moreover,in order to reduce controller updates,an event-triggered control law is developed for ADP algorithm.After that,to further simplify the complexity of controller design,this work is extended to a self-triggered case for relaxing the need for continuous signal monitoring by hardware devices.By employing the Lyapunov method,the uniform ultimate boundedness of all signals is proved to be guaranteed.Finally,a simulation example on a mass–spring–damper system subject to asymmetric input constraints is provided to validate the effectiveness of the proposed control scheme.

Decomposition of Mathematical Programming Models for Aircraft Wing Design Facilitating the Use of Dynamic Programming Approach

Aircraft designers strive to achieve optimal weight-reliability tradeoffs while designing an aircraft. Since aircraft wing skins account for more than fifty percent of their structural weight, aircraft wings must be designed with utmost care and attention in terms of material types and thickness configurations. In particular, the selection of thickness at each location of the aircraft wing skin is the most consequential task for aircraft designers. To accomplish this, we present discrete mathematical programming models to obtain optimal thicknesses either to minimize weight or to maximize reliability. We present theoretical results for the decomposition of these discrete mathematical programming models to reduce computer memory requirements and facilitate the use of dynamic programming for design purposes. In particular, a decomposed version of the weight minimization problem is solved for an aircraft wing with thirty locations (or panels) and fourteen thickness choices for each location to yield an optimal minimum weight design.

PDP: Parallel Dynamic Programming

Deep reinforcement learning is a focus research area in artificial intelligence. The principle of optimality in dynamic programming is a key to the success of reinforcement learning methods. The principle of adaptive dynamic programming(ADP)is first presented instead of direct dynamic programming(DP),and the inherent relationship between ADP and deep reinforcement learning is developed. Next, analytics intelligence, as the necessary requirement, for the real reinforcement learning, is discussed. Finally, the principle of the parallel dynamic programming, which integrates dynamic programming and analytics intelligence, is presented as the future computational intelligence.

Optimal Constrained Self-learning Battery Sequential Management in Microgrid Via Adaptive Dynamic Programming

This paper concerns a novel optimal self-learning battery sequential control scheme for smart home energy systems.The main idea is to use the adaptive dynamic programming(ADP) technique to obtain the optimal battery sequential control iteratively. First, the battery energy management system model is established, where the power efficiency of the battery is considered. Next, considering the power constraints of the battery, a new non-quadratic form performance index function is established, which guarantees that the value of the iterative control law cannot exceed the maximum charging/discharging power of the battery to extend the service life of the battery.Then, the convergence properties of the iterative ADP algorithm are analyzed, which guarantees that the iterative value function and the iterative control law both reach the optimums. Finally,simulation and comparison results are given to illustrate the performance of the presented method.

Residential Energy Scheduling for Variable Weather Solar Energy Based on Adaptive Dynamic Programming

The residential energy scheduling of solar energy is an important research area of smart grid. On the demand side, factors such as household loads, storage batteries, the outside public utility grid and renewable energy resources, are combined together as a nonlinear, time-varying, indefinite and complex system, which is difficult to manage or optimize. Many nations have already applied the residential real-time pricing to balance the burden on their grid. In order to enhance electricity efficiency of the residential micro grid, this paper presents an action dependent heuristic dynamic programming(ADHDP) method to solve the residential energy scheduling problem. The highlights of this paper are listed below. First,the weather-type classification is adopted to establish three types of programming models based on the features of the solar energy. In addition, the priorities of different energy resources are set to reduce the loss of electrical energy transmissions.Second, three ADHDP-based neural networks, which can update themselves during applications, are designed to manage the flows of electricity. Third, simulation results show that the proposed scheduling method has effectively reduced the total electricity cost and improved load balancing process. The comparison with the particle swarm optimization algorithm further proves that the present method has a promising effect on energy management to save cost.

Approximate Dynamic Programming for Stochastic Resource Allocation Problems

A stochastic resource allocation model, based on the principles of Markov decision processes(MDPs), is proposed in this paper. In particular, a general-purpose framework is developed, which takes into account resource requests for both instant and future needs. The considered framework can handle two types of reservations(i.e., specified and unspecified time interval reservation requests), and implement an overbooking business strategy to further increase business revenues. The resulting dynamic pricing problems can be regarded as sequential decision-making problems under uncertainty, which is solved by means of stochastic dynamic programming(DP) based algorithms. In this regard, Bellman’s backward principle of optimality is exploited in order to provide all the implementation mechanisms for the proposed reservation pricing algorithm. The curse of dimensionality, as the inevitable issue of the DP both for instant resource requests and future resource reservations,occurs. In particular, an approximate dynamic programming(ADP) technique based on linear function approximations is applied to solve such scalability issues. Several examples are provided to show the effectiveness of the proposed approach.

Parallel Control for Optimal Tracking via Adaptive Dynamic Programming

This paper studies the problem of optimal parallel tracking control for continuous-time general nonlinear systems.Unlike existing optimal state feedback control,the control input of the optimal parallel control is introduced into the feedback system.However,due to the introduction of control input into the feedback system,the optimal state feedback control methods can not be applied directly.To address this problem,an augmented system and an augmented performance index function are proposed firstly.Thus,the general nonlinear system is transformed into an affine nonlinear system.The difference between the optimal parallel control and the optimal state feedback control is analyzed theoretically.It is proven that the optimal parallel control with the augmented performance index function can be seen as the suboptimal state feedback control with the traditional performance index function.Moreover,an adaptive dynamic programming(ADP)technique is utilized to implement the optimal parallel tracking control using a critic neural network(NN)to approximate the value function online.The stability analysis of the closed-loop system is performed using the Lyapunov theory,and the tracking error and NN weights errors are uniformly ultimately bounded(UUB).Also,the optimal parallel controller guarantees the continuity of the control input under the circumstance that there are finite jump discontinuities in the reference signals.Finally,the effectiveness of the developed optimal parallel control method is verified in two cases.

Using approximate dynamic programming for multi-ESM scheduling to track ground moving targets

This paper researches the adaptive scheduling problem of multiple electronic support measures(multi-ESM) in a ground moving radar targets tracking application. It is a sequential decision-making problem in uncertain environment. For adaptive selection of appropriate ESMs, we generalize an approximate dynamic programming(ADP) framework to the dynamic case. We define the environment model and agent model, respectively. To handle the partially observable challenge, we apply the unsented Kalman filter(UKF) algorithm for belief state estimation. To reduce the computational burden, a simulation-based approach rollout with a redesigned base policy is proposed to approximate the long-term cumulative reward. Meanwhile, Monte Carlo sampling is combined into the rollout to estimate the expectation of the rewards. The experiments indicate that our method outperforms other strategies due to its better performance in larger-scale problems.

UAV flight strategy algorithm based on dynamic programming

Unmanned aerial vehicles(UAVs) may play an important role in data collection and offloading in vast areas deploying wireless sensor networks, and the UAV’s action strategy has a vital influence on achieving applicability and computational complexity. Dynamic programming(DP) has a good application in the path planning of UAV, but there are problems in the applicability of special terrain environment and the complexity of the algorithm.Based on the analysis of DP, this paper proposes a hierarchical directional DP(DDP) algorithm based on direction determination and hierarchical model. We compare our methods with Q-learning and DP algorithm by experiments, and the results show that our method can improve the terrain applicability, meanwhile greatly reduce the computational complexity.

Dynamic programming methodology for multi-criteria group decision-making under ordinal preferences

A method of minimizing rankings inconsistency is proposed for a decision-making problem with rankings of alternatives given by multiple decision makers according to multiple criteria. For each criteria, at first, the total inconsistency between the rankings of all alternatives for the group and the ones for every decision maker is defined after the decision maker weights in respect to the criteria are considered. Similarly, the total inconsistency between their final rankings for the group and the ones under every criteria is determined after the criteria weights are taken into account. Then two nonlinear integer programming models minimizing respectively the two total inconsistencies above are developed and then transformed to two dynamic programming models to obtain separately the rankings of all alternatives for the group with respect to each criteria and their final rankings. A supplier selection case illustrated the proposed method, and some discussions on the results verified its effectiveness. This work develops a new measurement of ordinal preferences’ inconsistency in multi-criteria group decision-making （MCGDM） and extends the cook-seiford social selection function to MCGDM considering weights of criteria and decision makers and can obtain unique ranking result.

CONTROLLING ROBOT MANIPULATORS BY DYNAMIC PROGRAMMING

A certain number of considerations should be taken into account in the dynamic control of robot manipulators as highly complex non-linear systems.In this article,we provide a detailed presentation of the mechanical and electrical impli- cations of robots equipped with DC motor actuators.This model takes into account all non-linear aspects of the system.Then,we develop computational algorithms for optimal control based on dynamic programming.The robot's trajectory must be predefined,but performance criteria and constraints applying to the system are not limited and we may adapt them freely to the robot and the task being studied.As an example,a manipulator arm with 3 degrees of freedom is analyzed.

Novel algorithm for distributed replicas management based on dynamic programming

Replicas can improve the data reliability in distributed system. However, the traditional algorithms for replica management are based on the assumption that all replicas have the uniform reliability, which is inaccurate in some actual systems. To address such problem, a novel algorithm is proposed based on dynamic programming to manage the number and distribution of replicas in different nodes. By using Markov model, replicas management is organized as a multi-phase process, and the recursion equations are provided. In this algorithm, the heterogeneity of nodes, the expense for maintaining replicas and the engaged space have been considered. Under these restricted conditions, this algorithm realizes high data reliability in a distributed system. The results of case analysis prove the feasibility of the algorithm.

Multi-view video color correction using dynamic programming

Color inconsistency between views is an important problem to be solved in multi-view video systems. A multi-view video color correction method using dynamic programming is proposed. Three-dimensional histograms are constructed with sequential conditional probability in HSI color space. Then, dynamic programming is used to seek the best color mapping relation with the minimum cost path between target image histogram and source image histogram. Finally, video tracking technique is performed to correct multi-view video. Experimental results show that the proposed method can obtain better subjective and objective performance in color correction.

Hierarchical adaptive stereo matching algorithm for obstacle detection with dynamic programming

An adaptive weighted stereo matching algorithm with multilevel and bidirectional dynamic programming based on ground control points （GCPs） is presented. To decrease time complexity without losing matching precision, using a multilevel search scheme, the coarse matching is processed in typical disparity space image, while the fine matching is processed in disparity-offset space image. In the upper level, GCPs are obtained by enhanced volumetric iterative algorithm enforcing the mutual constraint and the threshold constraint. Under the supervision of the highly reliable GCPs, bidirectional dynamic programming framework is employed to solve the inconsistency in the optimization path. In the lower level, to reduce running time, disparity-offset space is proposed to efficiently achieve the dense disparity image. In addition, an adaptive dual support-weight strategy is presented to aggregate matching cost, which considers photometric and geometric information. Further, post-processing algorithm can ameliorate disparity results in areas with depth discontinuities and related by occlusions using dual threshold algorithm, where missing stereo information is substituted from surrounding regions. To demonstrate the effectiveness of the algorithm, we present the two groups of experimental results for four widely used standard stereo data sets, including discussion on performance and comparison with other methods, which show that the algorithm has not only a fast speed, but also significantly improves the efficiency of holistic optimization.

Policy iteration optimal tracking control for chaotic systems by using an adaptive dynamic programming approach

A policy iteration algorithm of adaptive dynamic programming（ADP） is developed to solve the optimal tracking control for a class of discrete-time chaotic systems. By system transformations, the optimal tracking problem is transformed into an optimal regulation one. The policy iteration algorithm for discrete-time chaotic systems is first described. Then,the convergence and admissibility properties of the developed policy iteration algorithm are presented, which show that the transformed chaotic system can be stabilized under an arbitrary iterative control law and the iterative performance index function simultaneously converges to the optimum. By implementing the policy iteration algorithm via neural networks,the developed optimal tracking control scheme for chaotic systems is verified by a simulation.

Powertrain Fuel Consumption Modeling and Benchmark Analysis of a Parallel P4 Hybrid Electric Vehicle Using Dynamic Programming

The goal of this work is to develop a hybrid electric vehicle model that is suitable for use in a dynamic programming algorithm that provides the benchmark for optimal control of the hybrid powertrain. The benchmark analysis employs dynamic programming by backward induction to determine the globally optimal solution by solving the energy management problem starting at the final timestep and proceeding backwards in time. This method requires the development of a backwards facing model that propagates the wheel speed of the vehicle for the given drive cycle through the driveline components to determine the operating points of the powertrain. Although dynamic programming only searches the solution space within the feasible regions of operation, the benchmarking model must be solved for every admissible state at every timestep leading to strict requirements for runtime and memory. The backward facing model employs the quasi-static assumption of powertrain operation to reduce the fidelity of the model to accommodate these requirements. Verification and validation testing of the dynamic programming algorithm is conducted to ensure successful operation of the algorithm and to assess the validity of the determined control policy against a high-fidelity forward-facing vehicle model with a percent difference of fuel consumption of 1.2%. The benchmark analysis is conducted over multiple drive cycles to determine the optimal control policy that provides a benchmark for real-time algorithm development and determines control trends that can be used to improve existing algorithms. The optimal combined charge sustaining fuel economy of the vehicle is determined by the dynamic programming algorithm to be 32.99 MPG, a 52.6% increase over the stock 3.6 L 2019 Chevrolet Blazer.

Comparison between dynamic programming and genetic algorithm for hydro unit economic load dispatch

The hydro unit economic load dispatch （ELD） is of great importance in energy conservation and emission reduction. Dynamic programming （DP） and genetic algorithm （GA） are two representative algorithms for solving ELD problems. The goal of this study was to examine the performance of DP and GA while they were applied to ELD. We established numerical experiments to conduct performance comparisons between DP and GA with two given schemes. The schemes included comparing the CPU time of the algorithms when they had the same solution quality, and comparing the solution quality when they had the same CPU time. The numerical experiments were applied to the Three Gorges Reservoir in China, which is equipped with 26 hydro generation units. We found the relation between the performance of algorithms and the number of units through experiments. Results show that GA is adept at searching for optimal solutions in low-dimensional cases. In some cases, such as with a number of units of less than 10, GA＇s performance is superior to that of a coarse-grid DP. However, GA loses its superiority in high-dimensional cases. DP is powerful in obtaining stable and high-quality solutions. Its performance can be maintained even while searching over a large solution space. Nevertheless, due to its exhaustive enumerating nature, it costs excess time in low-dimensional cases.

A novel stable value iteration-based approximate dynamic programming algorithm for discrete-time nonlinear systems

The convergence and stability of a value-iteration-based adaptive dynamic programming （ADP） algorithm are con- sidered for discrete-time nonlinear systems accompanied by a discounted quadric performance index. More importantly than sufficing to achieve a good approximate structure, the iterative feedback control law must guarantee the closed-loop stability. Specifically, it is firstly proved that the iterative value function sequence will precisely converge to the optimum. Secondly, the necessary and sufficient condition of the optimal value function serving as a Lyapunov function is investi- gated. We prove that for the case of infinite horizon, there exists a finite horizon length of which the iterative feedback control law will provide stability, and this increases the practicability of the proposed value iteration algorithm. Neural networks （NNs） are employed to approximate the value functions and the optimal feedback control laws, and the approach allows the implementation of the algorithm without knowing the internal dynamics of the system. Finally, a simulation example is employed to demonstrate the effectiveness of the developed optimal control method.