Search

Power electronic converters are gradually becoming the main components of modern power systems due to the increasing integration of renewable energy sources. However, power converters may become unstable when interacting with the complex and time-varying power grid. To deal with this problem, an adaptive control design scheme for power converters is preferable, which can capture the closed-loop dynamical interaction between the converter and the grid via online data. In this paper, we propose an adaptive data-driven control method, called data-enabled policy optimization (DeePO), to stabilize power converters by using only online input-output data. Our contributions are threefold. First, we propose a covariance parameterization of partially observed linear systems with input-output data. Second, we develop a DeePO algorithm, which updates the parameterized policy with data-based gradient descent to achieve computationally efficient adaptive control. Third, we use high-fidelity simulations to verify that DeePO can effectively stabilize grid-connected power converters and quickly adapt to the changes in the power grid.

Data-enabled policy optimization (DeePO) is a newly proposed method to attack the open problem of direct adaptive LQR. In this work, we extend the DeePO framework to the linear quadratic tracking (LQT) with offline data. By introducing a covariance parameterization of the LQT policy, we derive a direct data-driven formulation of the LQT problem. Then, we use gradient descent method to iteratively update the parameterized policy to find an optimal LQT policy. Moreover, by revealing the connection between DeePO and model-based policy optimization, we prove the linear convergence of the DeePO iteration. Finally, a numerical experiment is given to validate the convergence results. We hope our work paves the way to direct adaptive LQT with online closed-loop data., Comment: 6 pages, 1 figures, submitted to ACC 2025

Learning policies in an asynchronous parallel way is essential to the numerous successes of RL for solving large-scale problems. However, their convergence performance is still not rigorously evaluated. To this end, we adopt the asynchronous parallel zero-order policy gradient (AZOPG) method to solve the continuous-time linear quadratic regulation problem. Specifically, as in the celebrated A3C algorithm, there are multiple parallel workers to asynchronously estimate PGs which are then sent to a central master for policy updates. Via quantifying its convergence rate of policy iterations, we show the linear speedup property of the AZOPG, both in theory and simulation, which clearly reveals the advantages of using parallel workers for learning policies., Comment: This article was submitted to IEEE TAC on Jan. 10, 2024

In safety-critical applications, reinforcement learning (RL) needs to consider safety constraints. However, theoretical understandings of constrained RL for continuous control are largely absent. As a case study, this paper presents a cost-constrained LQR formulation, where a number of LQR costs with user-defined penalty matrices are subject to constraints. To solve it, we propose a policy gradient primal-dual method to find an optimal state feedback gain. Despite the non-convexity of the cost-constrained LQR problem, we provide a constructive proof for strong duality and a geometric interpretation of an optimal multiplier set. By proving that the concave dual function is Lipschitz smooth, we further provide convergence guarantees for the PG primal-dual method. Finally, we perform simulations to validate our theoretical findings.

The traveling purchaser problem (TPP) is an important combinatorial optimization problem with broad applications. Due to the coupling between routing and purchasing, existing works on TPPs commonly address route construction and purchase planning simultaneously, which, however, leads to exact methods with high computational cost and heuristics with sophisticated design but limited performance. In sharp contrast, we propose a novel approach based on deep reinforcement learning (DRL), which addresses route construction and purchase planning separately, while evaluating and optimizing the solution from a global perspective. The key components of our approach include a bipartite graph representation for TPPs to capture the market-product relations, and a policy network that extracts information from the bipartite graph and uses it to sequentially construct the route. One significant benefit of our framework is that we can efficiently construct the route using the policy network, and once the route is determined, the associated purchasing plan can be easily derived through linear programming, while, leveraging DRL, we can train the policy network to optimize the global solution objective. Furthermore, by introducing a meta-learning strategy, the policy network can be trained stably on large-sized TPP instances, and generalize well across instances of varying sizes and distributions, even to much larger instances that are never seen during training. Experiments on various synthetic TPP instances and the TPPLIB benchmark demonstrate that our DRL-based approach can significantly outperform well-established TPP heuristics, reducing the optimality gap by 40%-90%, and also showing an advantage in runtime, especially on large-sized instances.

Model predictive control (MPC) solves a receding-horizon optimization problem in real-time, which can be computationally demanding when there are thousands of constraints. To accelerate online computation of MPC, we utilize data to adaptively remove the constraints while maintaining the MPC policy unchanged. Specifically, we design the removal rule based on the Lipschitz continuity of the MPC policy. This removal rule can use the information of historical data according to the Lipschitz constant and the distance between the current state and historical states. In particular, we provide the explicit expression for calculating the Lipschitz constant by the model parameters. Finally, simulations are performed to validate the effectiveness of the proposed method.

Direct data-driven design methods for the linear quadratic regulator (LQR) mainly use offline or episodic data batches, and their online adaptation has been acknowledged as an open problem. In this paper, we propose a direct adaptive method to learn the LQR from online closed-loop data. First, we propose a new policy parameterization based on the sample covariance to formulate a direct data-driven LQR problem, which is shown to be equivalent to the certainty-equivalence LQR with optimal non-asymptotic guarantees. Second, we design a novel data-enabled policy optimization (DeePO) method to directly update the policy, where the gradient is explicitly computed using only a batch of persistently exciting (PE) data. Third, we establish its global convergence via a projected gradient dominance property. Importantly, we efficiently use DeePO to adaptively learn the LQR by performing only one-step projected gradient descent per sample of the closed-loop system, which also leads to an explicit recursive update of the policy. Under PE inputs and for bounded noise, we show that the average regret of the LQR cost is upper-bounded by two terms signifying a sublinear decrease in time $\mathcal{O}(1/\sqrt{T})$ plus a bias scaling inversely with signal-to-noise ratio (SNR), which are independent of the noise statistics. Finally, we perform simulations to validate the theoretical results and demonstrate the computational and sample efficiency of our method., Comment: Submitted to IEEE TAC

Policy optimization (PO), an essential approach of reinforcement learning for a broad range of system classes, requires significantly more system data than indirect (identification-followed-by-control) methods or behavioral-based direct methods even in the simplest linear quadratic regulator (LQR) problem. In this paper, we take an initial step towards bridging this gap by proposing the data-enabled policy optimization (DeePO) method, which requires only a finite number of sufficiently exciting data to iteratively solve the LQR problem via PO. Based on a data-driven closed-loop parameterization, we are able to directly compute the policy gradient from a batch of persistently exciting data. Next, we show that the nonconvex PO problem satisfies a projected gradient dominance property by relating it to an equivalent convex program, leading to the global convergence of DeePO. Moreover, we apply regularization methods to enhance certainty-equivalence and robustness of the resulting controller and show an implicit regularization property. Finally, we perform simulations to validate our results., Comment: Accepted in IEEE CDC 2023

This work studies the standoff tracking problem to drive an unmanned aerial vehicle (UAV) to slide on a desired circle over a moving target at a constant height. We propose a novel Lyapunov guidance vector (LGV) field with tunable convergence rates for the UAV's trajectory planning and a deep neural network (DNN)-based model predictive control (MPC) scheme to track the reference trajectory. Then, we show how to collect samples for training the DNN offline and design an integral module (IM) to refine the tracking performance of our DNN-based MPC. Moreover, the hardware-in-the-loop (HIL) simulation with an FPGA@200MHz demonstrates that our method is a valid alternative to embedded implementations of MPC for addressing complex systems and applications which is impossible for directly solving the MPC optimization problems.

While the optimization landscape of policy gradient methods has been recently investigated for partially observed linear systems in terms of both static output feedback and dynamical controllers, they only provide convergence guarantees to stationary points. In this paper, we propose a new policy parameterization for partially observed linear systems, using a past input-output trajectory of finite length as feedback. We show that the solution set to the parameterized optimization problem is a matrix space, which is invariant to similarity transformation. By proving a gradient dominance property, we show the global convergence of policy gradient methods. Moreover, we observe that the gradient is orthogonal to the solution set, revealing an explicit relation between the resulting solution and the initial policy. Finally, we perform simulations to validate our theoretical results., Comment: To appear at IFAC World Congress 2023

Recent years have witnessed a booming interest in data-driven control of dynamical systems. However, the implicit data-driven output predictors are vulnerable to uncertainty such as process disturbance and measurement noise, causing unreliable predictions and unexpected control actions. In this brief, we put forward a new data-driven approach to output prediction of stochastic linear time-invariant (LTI) systems. By utilizing the innovation form, the uncertainty in stochastic LTI systems is recast as innovations that can be readily estimated from input-output data without knowing system matrices. In this way, by applying the fundamental lemma to the innovation form, we propose a new innovation-based data-driven output predictor (OP) of stochastic LTI systems, which bypasses the need for identifying state-space matrices explicitly and building a state estimator. The boundedness of the second moment of prediction errors in closed-loop is established under mild conditions. The proposed data-driven OP can be integrated into optimal control design for better performance. Numerical simulations demonstrate the outperformance of the proposed innovation-based methods in output prediction and control design over existing formulations., Comment: Accepted by Automatica

We propose and examine two optimal $(0,1)$-matrix completion problems with majorization ordered objectives. They elevate the seminal study by Gale and Ryser from feasibility to optimality in partial order programming (POP), referring to optimization with partially ordered objectives. We showcase their applications in electric vehicle charging, portfolio optimization, and secure data storage. Solving such integer POP (iPOP) problems is challenging because of the possible non-comparability among objective values and the integer requirements. Nevertheless, we prove the essential uniqueness of all optimal objective values and identify two particular ones for each of the two inherently symmetric iPOP problems. Furthermore, for every optimal objective value, we decompose the construction of an associated optimal~$(0,1)$-matrix into a series of sorting processes, respectively agreeing with the rule of thumb "peak shaving" or "valley filling." We show that the resulting algorithms have linear time complexities and verify their empirical efficiency via numerical simulations compared to the standard order-preserving method for POP., Comment: 16pages, 6 figures

In a direct data-driven approach, this paper studies the {\em property identification(ID)} problem to analyze whether an unknown linear system has a property of interest, e.g., stabilizability and structural properties. In sharp contrast to the model-based analysis, we approach it by directly using the input and state feedback data of the unknown system. Via a new concept of sufficient richness of input sectional data, we first establish the necessary and sufficient condition for the minimum input design to excite the system for property ID. Specifically, the input sectional data is sufficiently rich for property ID {\em if and only if} it spans a linear subspace that contains a property dependent minimum linear subspace, any basis of which can also be easily used to form the minimum excitation input. Interestingly, we show that many structural properties can be identified with the minimum input that is however unable to identify the explicit system model. Overall, our results rigorously quantify the advantages of the direct data-driven analysis over the model-based analysis for linear systems in terms of data efficiency.

System stabilization via policy gradient (PG) methods has drawn increasing attention in both control and machine learning communities. In this paper, we study their convergence and sample complexity for stabilizing linear time-invariant systems in terms of the number of system rollouts. Our analysis is built upon a discounted linear quadratic regulator (LQR) method which alternatively updates the policy and the discount factor of the LQR problem. Firstly, we propose an explicit rule to adaptively adjust the discount factor by exploring the stability margin of a linear control policy. Then, we establish the sample complexity of PG methods for stabilization, which only adds a coefficient logarithmic in the spectral radius of the state matrix to that for solving the LQR problem with a prior stabilizing policy. Finally, we perform simulations to validate our theoretical findings and demonstrate the effectiveness of our method on a class of nonlinear systems.

Control using quantized feedback is a fundamental approach to system synthesis with limited communication capacity. In this paper, we address the stabilization problem for unknown linear systems with logarithmically quantized feedback, via a direct data-driven control method. By leveraging a recently developed matrix S-lemma, we prove a sufficient and necessary condition for the existence of a common stabilizing controller for all possible dynamics consistent with data, in the form of a linear matrix inequality. Moreover, we formulate semi-definite programming to solve the coarsest quantization density. By establishing its connections to unstable eigenvalues of the state matrix, we further prove a necessary rank condition on the data for quantized feedback stabilization. Finally, we validate our theoretical results by numerical examples., Comment: To appear at the 4th Annual Conference on Learning for Dynamics and Control

This paper studies a distributed online constrained optimization problem over time-varying unbalanced digraphs without explicit subgradients. In sharp contrast to the existing algorithms, we design a novel consensus-based distributed online algorithm with a local randomized zeroth-order oracle and then rescale the oracle by constructing row-stochastic matrices, which aims to address the unbalancedness of time-varying digraphs. Under mild conditions, the average dynamic regret over a time horizon is shown to asymptotically converge at a sublinear rate provided that the accumulated variation grows sublinearly with a specific order. Moreover, the counterpart of the proposed algorithm when subgradients are available is also provided, along with its dynamic regret bound, which reflects that the convergence of our algorithm is essentially not affected by the zeroth-order oracle. Simulations on distributed targets tracking problem and dynamic sparse signal recovery problem in sensor networks are employed to demonstrate the effectiveness of the proposed algorithm., Comment: 13 pages, 8 figures

The package shipment problem requires to optimally co-design paths for both packages and a heterogeneous fleet in a transit center network (TCN). Instances arising from the package delivery industry in China usually involve more than ten thousand origin-destination (OD) pairs and have to be solved daily within an hour. Motivated by the fact that there is no interaction among different origin centers due to their competitive relationship, we propose a novel two-layer localized package shipment on a TCN (LPS-TCN) model that exploits outsourcing for cost saving. Consequently, the original problem breaks into a set of much smaller shipment problems, each of which has hundreds of OD pairs and is subsequently modelled as a mixed integer program (MIP). Since the LPS-TCN model is proved to be Strongly NP-hard and contains tens of thousands of feasible paths, an off-the-shelf MIP solver cannot produce a reliable solution in a practically acceptable amount of time. We develop a column generation based algorithm that iteratively adds "profitable" paths and further enhance it by problem-specific cutting planes and variable bound tightening techniques. Computational experiments on realistic instances from a major Chinese package express company demonstrate that the LPS-TCN model can yield solutions that bring daily economic cost reduction up to 1 million CNY for the whole TCN. In addition, our proposed algorithm solves the LPS-TCN model substantially faster than CPLEX, one of the state-of-the-art commercial MIP solvers.

Stability is one of the most fundamental requirements for systems synthesis. In this paper, we address the stabilization problem for unknown linear systems via policy gradient (PG) methods. We leverage a key feature of PG for Linear Quadratic Regulator (LQR), i.e., it drives the policy away from the boundary of the unstabilizing region along the descent direction, provided with an initial policy with finite cost. To this end, we discount the LQR cost with a factor, by adaptively increasing which gradient leads the policy to the stabilizing set while maintaining a finite cost. Based on the Lyapunov theory, we design an update rule for the discount factor which can be directly computed from data, rendering our method purely model-free. Compared to recent work \citep{perdomo2021stabilizing}, our algorithm allows the policy to be updated only once for each discount factor. Moreover, the number of sampled trajectories and simulation time for gradient descent is significantly reduced to $\mathcal{O}(\log(1/\epsilon))$ for the desired accuracy $\epsilon$. Finally, we conduct simulations on both small-scale and large-scale examples to show the efficiency of our discount PG method., Comment: Submitted to L4DC 2022

The key of the post-disaster humanitarian logistics (PD-HL) is to build a good facility location and capacity planning (FLCP) model for delivering relief supplies to affected areas in time. To fully exploit the historical PD data, this paper adopts the data-driven distributionally robust (DR) approach and proposes a novel multi-period FLCP model under the $\infty$-Wasserstein joint chance constraints (MFLCP-W). Specifically, we sequentially decide locations from a candidate set to build facilities with supply capacities, which are expanded if more economical, and use a finite number of historical demand samples in chance constraints to ensure a high probability of on-time delivery. To solve the MFLCP-W model, we equivalently reformulate it as a mixed integer second-order cone program and then solve it by designing an effective outer approximation algorithm with two tailored valid cuts. Finally, a case study under hurricane threats shows that MFLCP-W outperforms its counterparts in the terms of the cost and service quality, and that our algorithm converges significantly faster than the commercial solver CPLEX 12.8 with a better optimality gap.

Information compression is essential to reduce communication cost in distributed optimization over peer-to-peer networks. This paper proposes a communication-efficient linearly convergent distributed (COLD) algorithm to solve strongly convex optimization problems. By compressing innovation vectors, which are the differences between decision vectors and their estimates, COLD is able to achieve linear convergence for a class of $\delta$-contracted compressors. We explicitly quantify how the compression affects the convergence rate and show that COLD matches the same rate of its uncompressed version. To accommodate a wider class of compressors that includes the binary quantizer, we further design a novel dynamical scaling mechanism and obtain the linearly convergent Dyna-COLD. Importantly, our results strictly improve existing results for the quantized consensus problem. Numerical experiments demonstrate the advantages of both algorithms under different compressors., Comment: 14 pages

This paper considers a networked aggregative game (NAG) where the players are distributed over a communication network. By only communicating with a subset of players, the goal of each player in the NAG is to minimize an individual cost function that depends on its own action and the aggregate of all the players' actions. To this end, we design a novel distributed algorithm that jointly exploits the ideas of the consensus algorithm and the conditional projection descent. Under strongly monotone assumption on the pseudo-gradient mapping, the proposed algorithm with fixed step-sizes is proved to converge linearly to the unique Nash equilibrium of the NAG. Then the theoretical results are validated by numerical experiments., Comment: 6 pages, 2 figures, submitted to the IEEE CDC 2021

While the techniques in optimal control theory are often model-based, the policy optimization (PO) approach directly optimizes the performance metric of interest. Even though it has been an essential approach for reinforcement learning problems, there is little theoretical understanding on its performance. In this paper, we focus on the risk-constrained linear quadratic regulator (RC-LQR) problem via the PO approach, which requires addressing a challenging non-convex constrained optimization problem. To solve it, we first build on our earlier result that an optimal policy has a time-invariant affine structure to show that the associated Lagrangian function is coercive, locally gradient dominated and has local Lipschitz continuous gradient, based on which we establish strong duality. Then, we design policy gradient primal-dual methods with global convergence guarantees in both model-based and sample-based settings. Finally, we use samples of system trajectories in simulations to validate our methods.

Communication efficiency is a major bottleneck in the applications of distributed networks. To address the problem, the problem of quantized distributed optimization has attracted a lot of attention. However, most of the existing quantized distributed optimization algorithms can only converge sublinearly. To achieve linear convergence, this paper proposes a novel quantized distributed gradient tracking algorithm (Q-DGT) to minimize a finite sum of local objective functions over directed networks. Moreover, we explicitly derive the update rule for the number of quantization levels, and prove that Q-DGT can converge linearly even when the exchanged variables are respectively one bit. Numerical results also confirm the efficiency of the proposed algorithm., Comment: Accepted by IEEE Transactions on Automatic Control as a technical note

The behaviour of a stochastic dynamical system may be largely influenced by those low-probability, yet extreme events. To address such occurrences, this paper proposes an infinite-horizon risk-constrained Linear Quadratic Regulator (LQR) framework with time-average cost. In addition to the standard LQR objective, the average one-stage predictive variance of the state penalty is constrained to lie within a user-specified level. By leveraging the duality, its optimal solution is first shown to be stationary and affine in the state, i.e., $u(x,\lambda^*) = -K(\lambda^*)x + l(\lambda^*)$, where $\lambda^*$ is an optimal multiplier, used to address the risk constraint. Then, we establish the stability of the resulting closed-loop system. Furthermore, we propose a primal-dual method with sublinear convergence rate to find an optimal policy $u(x,\lambda^*)$. Finally, a numerical example is provided to demonstrate the effectiveness of the proposed framework and the primal-dual method., Comment: Submitted to IEEE CDC 2021

This paper studies the distributed state estimation in sensor network, where $m$ sensors are deployed to infer the $n$-dimensional state of a linear time-invariant (LTI) Gaussian system. By a lossless decomposition of optimal steady-state Kalman filter, we show that the problem of distributed estimation can be reformulated as synchronization of homogeneous linear systems. Based on such decomposition, a distributed estimator is proposed, where each sensor node runs a local filter using only its own measurement and fuses the local estimate of each node with a consensus algorithm. We show that the average of the estimate from all sensors coincides with the optimal Kalman estimate. Numerical examples are provided in the end to illustrate the performance of the proposed scheme.

We propose a fully asynchronous networked aggregative game (Asy-NAG) where each player minimizes a cost function that depends on its local action and the aggregate of all players' actions. In sharp contrast to the existing NAGs, each player in our Asy-NAG can compute an estimate of the aggregate action at any wall-clock time by only using (possibly stale) information from nearby players of a directed network. Such an asynchronous update does not require any coordination among players. Moreover, we design a novel distributed algorithm with an aggressive mechanism for each player to adaptively adjust the optimization stepsize per update. Particularly, the slow players in terms of updating their estimates smartly increase their stepsizes to catch up with the fast ones. Then, we develop an augmented system approach to address the asynchronicity and the information delays between players, and rigorously show the convergence to a Nash equilibrium of the Asy-NAG via a perturbed coordinate algorithm which is also of independent interest. Finally, we evaluate the performance of the distributed algorithm through numerical simulations., Comment: 19 pages, 11 figures

Risk-aware control, though with promise to tackle unexpected events, requires a known exact dynamical model. In this work, we propose a model-free framework to learn a risk-aware controller with a focus on the linear system. We formulate it as a discrete-time infinite-horizon LQR problem with a state predictive variance constraint. To solve it, we parameterize the policy with a feedback gain pair and leverage primal-dual methods to optimize it by solely using data. We first study the optimization landscape of the Lagrangian function and establish the strong duality in spite of its non-convex nature. Alongside, we find that the Lagrangian function enjoys an important local gradient dominance property, which is then exploited to develop a convergent random search algorithm to learn the dual function. Furthermore, we propose a primal-dual algorithm with global convergence to learn the optimal policy-multiplier pair. Finally, we validate our results via simulations., Comment: To appear in the Annual Conference on Learning for Dynamics and Control (L4DC) 2021

Stochastic optimal control usually requires an explicit dynamical model with probability distributions, which are difficult to obtain in practice. In this work, we consider the linear quadratic regulator (LQR) problem of unknown linear systems and adopt a Wasserstein penalty to address the distribution uncertainty of additive stochastic disturbances. By constructing an equivalent deterministic game of the penalized LQR problem, we propose a Q-learning method with convergence guarantees to learn an optimal minimax controller., Comment: Accepted by Automatica, to appear in 2023

The isoline tracking of this work is concerned with the control design for a sensing vehicle to track a desired isoline of an unknown scalar field. To this end, we propose a simple PI-like controller for a Dubins vehicle in the GPS-denied environments. Our key idea lies in the design of a novel sliding surface based error in the standard PI controller. For the circular field, we show that the P-like controller can globally regulate the vehicle to the desired isoline with the steady-state error that can be arbitrarily reduced by increasing the P gain, and is eliminated by the PI-like controller. For any smoothing field, the P-like controller is able to achieve the local regulation. Then, it is extended to the cases of a single-integrator vehicle and a doubleintegrator vehicle, respectively. Finally, the effectiveness and advantages of our approaches are validated via simulations on the fixed-wing UAV and quadrotor simulators.

The isoline tracking of this work is concerned with the control design for a sensing robot to track a given isoline of an unknown 2-D scalar filed. To this end, we propose a coordinate-free controller with a simple PI-like form using only the concentration feedback for a Dubins robot, which is particularly useful in GPS-denied environments. The key idea lies in the novel design of a sliding surface based error term in the standard PI controller. Interestingly, we also prove that the tracking error can be reduced by increasing the proportion gain, and is eliminated for circular fields with a non-zero integral gain. The effectiveness of our controller is validated via simulations by using a fixed-wing UAV on the real dataset of the concentration distribution of PM 2.5 in Handan, China., Comment: 6 pages, 3 figures

During recent decades, the automatic train operation (ATO) system has been gradually adopted in many subway systems for its low-cost and intelligence. This paper proposes two smart train operation algorithms by integrating the expert knowledge with reinforcement learning algorithms. Compared with previous works, the proposed algorithms can realize the control of continuous action for the subway system and optimize multiple critical objectives without using an offline speed profile. Firstly, through learning historical data of experienced subway drivers, we extract the expert knowledge rules and build inference methods to guarantee the riding comfort, the punctuality, and the safety of the subway system. Then we develop two algorithms for optimizing the energy efficiency of train operation. One is the smart train operation (STO) algorithm based on deep deterministic policy gradient named (STOD) and the other is the smart train operation algorithm based on normalized advantage function (STON). Finally, we verify the performance of proposed algorithms via some numerical simulations with the real field data from the Yizhuang Line of the Beijing Subway and illustrate that the developed smart train operation algorithm are better than expert manual driving and existing ATO algorithms in terms of energy efficiency. Moreover, STOD and STON can adapt to different trip times and different resistance conditions.

This paper proposes a \emph{fully asynchronous} scheme for the policy evaluation problem of distributed reinforcement learning (DisRL) over directed peer-to-peer networks. Without waiting for any other node of the network, each node can locally update its value function at any time by using (possibly delayed) information from its neighbors. This is in sharp contrast to the gossip-based scheme where a pair of nodes concurrently update. Though the fully asynchronous setting involves a difficult multi-timescale decision problem, we design a novel stochastic average gradient (SAG) based distributed algorithm and develop a push-pull augmented graph approach to prove its exact convergence at a linear rate of $\mathcal{O}(c^k)$ where $c\in(0,1)$ and $k$ increases by one no matter on which node updates. Finally, numerical experiments validate that our method speeds up linearly with respect to the number of nodes, and is robust to straggler nodes.

This paper considers the distributed optimization problem where each node of a peer-to-peer network minimizes a finite sum of objective functions by communicating with its neighboring nodes. In sharp contrast to the existing literature where the fastest distributed algorithms converge either with a global linear or a local superlinear rate, we propose a distributed adaptive Newton (DAN) algorithm with a global quadratic convergence rate. Our key idea lies in the design of a finite-time set-consensus method with Polyak's adaptive stepsize. Moreover, we introduce a low-rank matrix approximation (LA) technique to compress the innovation of Hessian matrix so that each node only needs to transmit message of dimension $\mathcal{O}(p)$ (where $p$ is the dimension of decision vectors) per iteration, which is essentially the same as that of first-order methods. Nevertheless, the resulting DAN-LA converges to an optimal solution with a global superlinear rate. Numerical experiments on logistic regression problems are conducted to validate their advantages over existing methods., Comment: Accepted to Automatica as regular paper. 13 pages

This paper proposes a second-order conic programming (SOCP) approach to solve distributionally robust two-stage stochastic linear programs over 1-Wasserstein balls. We start from the case with distribution uncertainty only in the objective function and exactly reformulate it as an SOCP problem. Then, we study the case with distribution uncertainty only in constraints, and show that such a robust program is generally NP-hard as it involves a norm maximization problem over a polyhedron. However, it is reduced to an SOCP problem if the extreme points of the polyhedron are given as a prior. This motivates to design a constraint generation algorithm with provable convergence to approximately solve the NP-hard problem. In sharp contrast to the exiting literature, the distribution achieving the worst-case cost is given as an "empirical" distribution by simply perturbing each sample for both cases. Finally, experiments illustrate the advantages of the proposed model in terms of the out-of-sample performance and the computational complexity.

This paper proposes a coordinate-free controller for a nonholonomic vehicle to circumnavigate a fully-actuated moving target by using range-only measurements. If the range rate is available, our Proportional Derivative (PD)-like controller has a simple structure as the standard PD controller, except the design of an additive constant bias and a saturation function in the error feedback. We show that if the target is stationary, the vehicle asymptotically encloses the target with a predefined radius at an exponential convergence rate, i.e., an exact circumnavigation pattern can be completed. For a moving target, the circumnavigation error converges to a small region whose size is shown proportional to the maneuverability of the target, e.g., the maximum linear speed and acceleration. Moreover, we design a second-order sliding mode (SOSM) filter to estimate the range rate and show that the SOSM filter can recover the range rate in a finite time. Finally, the effectiveness and advantages of our controller are validated via both numerical simulations and real experiments., Comment: 13 pages,17 figures

Recently, reinforcement learning (RL) has been extensively studied and achieved promising results in a wide range of control tasks. Meanwhile, autonomous underwater vehicle (AUV) is an important tool for executing complex and challenging underwater tasks. The advances in RL offers ample opportunities for developing intelligent AUVs. This paper provides a selected review on RL based control for AUVs with the focus on applications of RL to low-level control tasks for underwater regulation and tracking. To this end, we first present a concise introduction to the RL based control framework. Then, we provide an overview of RL methods for AUVs control problems, where the main challenges and recent progresses are discussed. Finally, two representative cases of RL-based controllers are given in detail for the model-free RL methods on AUVs.

The suspension regulation is critical to the operation of medium-low-speed maglev trains (mlsMTs). Due to uncertain environment, strong disturbances and high nonlinearity of the system dynamics, this problem cannot be well solved by most of the model-based controllers. In this paper, we propose a model-free controller by reformulating it as a continuous-state, continuous-action Markov decision process (MDP) with unknown transition probabilities. With the deterministic policy gradient and neural network approximation, we design reinforcement learning (RL) algorithms to solve the MDP and obtain a state-feedback controller by using sampled data from the suspension system. To further improve its performance, we adopt a double Q-learning scheme for learning the regulation controller. We illustrate that the proposed controllers outperform the existing PID controller with a real dataset from the mlsMT in Changsha, China and is even comparable to model-based controllers, which assume that the complete information of the model is known, via simulations., Comment: 12 pages, 15 figures

This paper proposes a distributed dual gradient tracking algorithm (DDGT) to solve resource allocation problems over an unbalanced network, where each node in the network holds a private cost function and computes the optimal resource by interacting only with its neighboring nodes. Our key idea is the novel use of the distributed push-pull gradient algorithm (PPG) to solve the dual problem of the resource allocation problem. To study the convergence of the DDGT, we first establish the sublinear convergence rate of PPG for non-convex objective functions, which advances the existing results on PPG as they require the strong-convexity of objective functions. Then we show that the DDGT converges linearly for strongly convex and Lipschitz smooth cost functions, and sublinearly without the Lipschitz smoothness. Finally, experimental results suggest that DDGT outperforms existing algorithms., Comment: Accepted by IEEE Transactions on Signal Processing. This version fixed some typos in the accepted version

This paper studies a decentralized stochastic gradient tracking (DSGT) algorithm for non-convex empirical risk minimization problems over a peer-to-peer network of nodes, which is in sharp contrast to the existing DSGT only for convex problems. To ensure exact convergence and handle the variance among decentralized datasets, each node performs a stochastic gradient (SG) tracking step by using a mini-batch of samples, where the batch size is designed to be proportional to the size of the local dataset. We explicitly evaluate the convergence rate of DSGT with respect to the number of iterations in terms of algebraic connectivity of the network, mini-batch size, gradient variance, etc. Under certain conditions, we further show that DSGT has a network independence property in the sense that the network topology only affects the convergence rate up to a constant factor. Hence, the convergence rate of DSGT can be comparable to the centralized SGD method. Moreover, a linear speedup of DSGT with respect to the number of nodes is achievable for some scenarios. Numerical experiments for neural networks and logistic regression problems on CIFAR-10 finally illustrate the advantages of DSGT., Comment: This paper has been revised and theoretical results are improved

This work aims to design an optimization-based controller for a discrete-time Dubins vehicle to approach a target with unknown position as fast as possible by only using bearing measurements. To this end, we propose a bi-objective optimization problem, which jointly considers the performance of estimating the unknown target position and controlling the mobile vehicle to a known position, and then adopt a weighted sum method with normalization to solve it. The controller is given based on the solution of the optimization problem in ties with a least-square estimate of the target position. Moreover, the controller does not need the vehicle's global position information. Finally, simulation results are included to validate the effectiveness of the proposed controller.

This paper proposes a coordinate-free controller to drive a mobile robot to encircle a target at unknown position by only using range measurements. Different from the existing works, a backstepping based controller is proposed to encircle the target with zero steady-state error for any desired smooth pattern. Moreover, we show its asymptotic exponential convergence under a fixed set of control parameters, which are independent of the initial distance to the target. The effectiveness and advantages of the proposed controller are validated via simulations.

This paper proposes a flight controller for an unmanned aerial vehicle (UAV) to loiter over a ground moving target (GMT). We are concerned with the scenario that the stochastically time-varying maneuver of the GMT is unknown to the UAV, which renders it challenging to estimate the GMT's motion state. Assuming that the state of the GMT is available, we first design a discrete-time Lyapunov vector field for the loitering guidance and then design a discrete-time integral sliding mode control (ISMC) to track the guidance commands. By modeling the maneuver process as a finite-state Markov chain, we propose a Rao-Blackwellised particle filter (RBPF), which only requires a few number of particles, to simultaneously estimate the motion state and the maneuver of the GMT with a camera or radar sensor. Then, we apply the principle of certainty equivalence to the ISMC and obtain the flight controller for completing the loitering task. Finally, the effectiveness and advantages of our controller are validated via simulations., Comment: 12 pages, 14 figures. Accepted by IEEE Transactions on Control Systems Technology

This paper proposes a data-driven distributionally robust shortest path (DRSP) model where the distribution of the travel time in the transportation network can only be partially observed through a finite number of samples. Specifically, we aim to find an optimal path to minimize the worst-case $\alpha$-reliable mean-excess travel time (METT) over a Wasserstein ball, which is centered at the empirical distribution of the sample dataset and the ball radius quantifies the level of its confidence. In sharp contrast to the existing DRSP models, our model is equivalently reformulated as a tractable mixed 0-1 convex problem, e.g., 0-1 linear program or 0-1 second-order cone program. Moreover, we also explicitly derive the distribution achieving the worst-case METT by simply perturbing each sample. Experiments demonstrate the advantages of our DRSP model in terms of the out-of-sample performance and computational complexity. Finally, our DRSP model is easily extended to solve the DR bi-criteria shortest path problem and the minimum cost flow problem.

We consider the distributed optimization problem, the goal of which is to minimize the sum of local objective functions over a directed network. Though it has been widely studied recently, most of the existing algorithms are designed for synchronized or randomly activated implementation, which may create deadlocks in practice. In sharp contrast, we propose a \emph{fully} asynchronous push-pull gradient algorithm (APPG) where each node updates without waiting for any other node by using (possibly stale) information from neighbors. Thus, it is both deadlock-free and robust to any bounded communication delay. Moreover, we construct two novel augmented networks to theoretically evaluate its performance from the worst-case point of view and show that if local functions have Lipschitz-continuous gradients and their sum satisfies the Polyak-\L ojasiewicz condition (convexity is not required), each node of APPG converges to the same optimal solution at a linear rate of $\mathcal{O}(\lambda^k)$, where $\lambda\in(0,1)$ and the virtual counter $k$ increases by one no matter which node updates. This largely elucidates its linear speedup efficiency and shows its advantage over the synchronous version. Finally, the performance of APPG is numerically validated via a logistic regression problem on the \emph{Covertype} dataset., Comment: 14 pages

This paper studies the cooperative source seeking problem via a networked multi-vehicle system. In contrast to existing literature, the multi-vehicle system is controlled to the source position that maximizes aggregated multiple unknown scalar fields and each sensor-enabled vehicle only samples measurements of one scalar field. Thus, a single vehicle is unable to localize the source and has to cooperate with its neighboring vehicles. By jointly exploiting the ideas of the consensus algorithm and the stochastic extremum seeking (ES), this paper proposes novel distributed stochastic ES controllers, which are gradient-free and do not need any absolute information, such that the multi-vehicle system simultaneously approaches the source position. The effectiveness of the proposed controllers is proved for quadratic scalar fields. Finally, illustrative examples are included to validate the theoretical results.