arXiv subjects

Sean P. Meyn

Publications and source records attributed to Sean P. Meyn.

At least 19 recordsLinked to original sources

arXiv2025

Interacting Particle Systems for Fast Linear Quadratic RL

This paper is concerned with the design of algorithms based on systems of interacting particles to represent, approximate, and learn the optimal control law for reinforcement learning (RL). The primary contribution is that convergence rates are greatly accelerated by the interactions between particles. Theory focuses on the linear quadratic stochastic optimal control problem for which a complete and novel theory is presented. Apart from the new algorithm, sample complexity bounds are obtained, and it is shown that the mean square error scales as $1/N$ where $N$ is the number of particles. The theoretical results and algorithms are illustrated with numerical experiments and comparisons with other recent approaches, where the faster convergence of the proposed algorithm is numerically demonstrated.

eess.SY↗

arXiv2024

Dual Ensemble Kalman Filter for Stochastic Optimal Control

In this paper, stochastic optimal control problems in continuous time and space are considered. In recent years, such problems have received renewed attention from the lens of reinforcement learning (RL) which is also one of our motivation. The main contribution is a simulation-based algorithm -- dual ensemble Kalman filter (EnKF) -- to numerically approximate the solution of these problems. The paper extends our previous work where the dual EnKF was applied in deterministic settings of the problem. The theoretical results and algorithms are illustrated with numerical experiments.

eess.SY↗

arXiv2022

Controlled Interacting Particle Algorithms for Simulation-based Reinforcement Learning

This paper is concerned with optimal control problems for control systems in continuous time, and interacting particle system methods designed to construct approximate control solutions. Particular attention is given to the linear quadratic (LQ) control problem. There is a growing interest in re-visiting this classical problem, in part due to the successes of reinforcement learning (RL). The main question of this body of research (and also of our paper) is to approximate the optimal control law {\em without} explicitly solving the Riccati equation. A novel simulation-based algorithm, namely a dual ensemble Kalman filter (EnKF), is introduced. The algorithm is used to obtain formulae for optimal control, expressed entirely in terms of the EnKF particles. An extension to the nonlinear case is also presented. The theoretical results and algorithms are illustrated with numerical experiments.

eess.SY↗

arXiv2020

Convex Q-Learning, Part 1: Deterministic Optimal Control

It is well known that the extension of Watkins' algorithm to general function approximation settings is challenging: does the projected Bellman equation have a solution? If so, is the solution useful in the sense of generating a good policy? And, if the preceding questions are answered in the affirmative, is the algorithm consistent? These questions are unanswered even in the special case of Q-function approximations that are linear in the parameter. The challenge seems paradoxical, given the long history of convex analytic approaches to dynamic programming. The paper begins with a brief survey of linear programming approaches to optimal control, leading to a particular over parameterization that lends itself to applications in reinforcement learning. The main conclusions are summarized as follows: (i) The new class of convex Q-learning algorithms is introduced based on the convex relaxation of the Bellman equation. Convergence is established under general conditions, including a linear function approximation for the Q-function. (ii) A batch implementation appears similar to the famed DQN algorithm (one engine behind AlphaZero). It is shown that in fact the algorithms are very different: while convex Q-learning solves a convex program that approximates the Bellman equation, theory for DQN is no stronger than for Watkins' algorithm with function approximation: (a) it is shown that both seek solutions to the same fixed point equation, and (b) the ODE approximations for the two algorithms coincide, and little is known about the stability of this ODE. These results are obtained for deterministic nonlinear systems with total cost criterion. Many extensions are proposed, including kernel implementation, and extension to MDP models.

math.OC↗

arXiv2020

Zap Q-Learning With Nonlinear Function Approximation

Zap Q-learning is a recent class of reinforcement learning algorithms, motivated primarily as a means to accelerate convergence. Stability theory has been absent outside of two restrictive classes: the tabular setting, and optimal stopping. This paper introduces a new framework for analysis of a more general class of recursive algorithms known as stochastic approximation. Based on this general theory, it is shown that Zap Q-learning is consistent under a non-degeneracy assumption, even when the function approximation architecture is nonlinear. Zap Q-learning with neural network function approximation emerges as a special case, and is tested on examples from OpenAI Gym. Based on multiple experiments with a range of neural network sizes, it is found that the new algorithms converge quickly and are robust to choice of function approximation architecture.

cs.LG↗

arXiv2020

Q-learning with Uniformly Bounded Variance: Large Discounting is Not a Barrier to Fast Learning

Sample complexity bounds are a common performance metric in the Reinforcement Learning literature. In the discounted cost, infinite horizon setting, all of the known bounds have a factor that is a polynomial in $1/(1-γ)$, where $γ< 1$ is the discount factor. For a large discount factor, these bounds seem to imply that a very large number of samples is required to achieve an $\varepsilon$-optimal policy. The objective of the present work is to introduce a new class of algorithms that have sample complexity uniformly bounded for all $γ< 1$. One may argue that this is impossible, due to a recent min-max lower bound. The explanation is that this previous lower bound is for a specific problem, which we modify, without compromising the ultimate objective of obtaining an $\varepsilon$-optimal policy. Specifically, we show that the asymptotic covariance of the Q-learning algorithm with an optimized step-size sequence is a quadratic function of $1/(1-γ)$; an expected, and essentially known result. The new relative Q-learning algorithm proposed here is shown to have asymptotic covariance that is a quadratic in $1/(1- ρ^* γ)$, where $1 - ρ^* > 0$ is an upper bound on the spectral gap of an optimal transition matrix.

cs.LG↗

arXiv2020

Variance Reduction in Simulation of Multiclass Processing Networks

We use simulation to estimate the steady-state performance of a stable multiclass queueing network. Standard estimators have been seen to perform poorly when the network is heavily loaded. We introduce two new simulation estimators. The first provides substantial variance reductions in moderately-loaded networks at very little additional computational cost. The second estimator provides substantial variance reductions in heavy traffic, again for a small additional computational cost. Both methods employ the variance reduction method of control variates, and differ in terms of how the control variates are constructed.

math.PR↗

arXiv2019

What is the Lagrangian for Nonlinear Filtering?

Duality between estimation and optimal control is a problem of rich historical significance. The first duality principle appears in the seminal paper of Kalman-Bucy, where the problem of minimum variance estimation is shown to be dual to a linear quadratic (LQ) optimal control problem. Duality offers a constructive proof technique to derive the Kalman filter equation from the optimal control solution. This paper generalizes the classical duality result of Kalman-Bucy to the nonlinear filter: The state evolves as a continuous-time Markov process and the observation is a nonlinear function of state corrupted by an additive Gaussian noise. A dual process is introduced as a backward stochastic differential equation (BSDE). The process is used to transform the problem of minimum variance estimation into an optimal control problem. Its solution is obtained from an application of the maximum principle, and subsequently used to derive the equation of the nonlinear filter. The classical duality result of Kalman-Bucy is shown to be a special case.

math.OC↗

arXiv2019

Zap Q-Learning for Optimal Stopping Time Problems

The objective in this paper is to obtain fast converging reinforcement learning algorithms to approximate solutions to the problem of discounted cost optimal stopping in an irreducible, uniformly ergodic Markov chain, evolving on a compact subset of $\mathbb{R}^n$. We build on the dynamic programming approach taken by Tsitsikilis and Van Roy, wherein they propose a Q-learning algorithm to estimate the optimal state-action value function, which then defines an optimal stopping rule. We provide insights as to why the convergence rate of this algorithm can be slow, and propose a fast-converging alternative, the "Zap-Q-learning" algorithm, designed to achieve optimal rate of convergence. For the first time, we prove the convergence of the Zap-Q-learning algorithm under the assumption of linear function approximation setting. We use ODE analysis for the proof, and the optimal asymptotic variance property of the algorithm is reflected via fast convergence in a finance example.

eess.SY↗

arXiv2019

An Approach to Duality in Nonlinear Filtering

This paper revisits the question of duality between minimum variance estimation and optimal control first described for the linear Gaussian case in the celebrated paper of Kalman and Bucy. A duality result is established for nonlinear filtering, mirroring closely the original Kalman-Bucy duality of control and estimation for linear systems. The result for the finite state-space continuous time Markov chain is presented. It's solution is used to derive the classical Wonham filter.

math.PR↗

arXiv2019

Diffusion map-based algorithm for Gain function approximation in the Feedback Particle Filter

Feedback particle filter (FPF) is a numerical algorithm to approximate the solution of the nonlinear filtering problem in continuous-time settings. In any numerical implementation of the FPF algorithm, the main challenge is to numerically approximate the so-called gain function. A numerical algorithm for gain function approximation is the subject of this paper. The exact gain function is the solution of a Poisson equation involving a probability-weighted Laplacian $\Delta_\rho$. The numerical problem is to approximate this solution using {\em only} finitely many particles sampled from the probability distribution $\rho$. A diffusion map-based algorithm was proposed by the authors in a prior work to solve this problem. The algorithm is named as such because it involves, as an intermediate step, a diffusion map approximation of the exact semigroup $e^{\Delta_\rho}$. The original contribution of this paper is to carry out a rigorous error analysis of the diffusion map-based algorithm. The error is shown to include two components: bias and variance. The bias results from the diffusion map approximation of the exact semigroup. The variance arises because of finite sample size. Scalings and upper bounds are derived for bias and variance. These bounds are then illustrated with numerical experiments that serve to emphasize the effects of problem dimension and sample size. The proposed algorithm is applied to two filtering examples and comparisons provided with the sequential importance resampling (SIR) particle filter.

math.OC↗

arXiv2018

Differential Temporal Difference Learning

Value functions derived from Markov decision processes arise as a central component of algorithms as well as performance metrics in many statistics and engineering applications of machine learning techniques. Computation of the solution to the associated Bellman equations is challenging in most practical cases of interest. A popular class of approximation techniques, known as Temporal Difference (TD) learning algorithms, are an important sub-class of general reinforcement learning methods. The algorithms introduced in this paper are intended to resolve two well-known difficulties of TD-learning approaches: Their slow convergence due to very high variance, and the fact that, for the problem of computing the relative value function, consistent algorithms exist only in special cases. First we show that the gradients of these value functions admit a representation that lends itself to algorithm design. Based on this result, a new class of differential TD-learning algorithms is introduced. For Markovian models on Euclidean space with smooth dynamics, the algorithms are shown to be consistent under general conditions. Numerical results show dramatic variance reduction when compared to standard methods.

cs.LG↗

arXiv2017

Distributed computation of equilibria in misspecified convex stochastic Nash games

The distributed computation of Nash equilibria is assuming growing relevance in engineering where such problems emerge in the context of distributed control. Accordingly, we present schemes for computing equilibria of two classes of static stochastic convex games complicated by a parametric misspecification, a natural concern in the control of large-scale networked engineered system. In both schemes, players learn the equilibrium strategy while resolving the misspecification: (1) Monotone stochastic Nash games: We present a set of coupled stochastic approximation schemes distributed across agents in which the first scheme updates each agent's strategy via a projected (stochastic) gradient step while the second scheme updates every agent's belief regarding its misspecified parameter using an independently specified learning problem. We proceed to show that the produced sequences converge in an almost-sure sense to the true equilibrium strategy and the true parameter, respectively. Surprisingly, convergence in the equilibrium strategy achieves the optimal rate of convergence in a mean-squared sense with a quantifiable degradation in the rate constant; (2) Stochastic Nash-Cournot games with unobservable aggregate output: We refine (1) to a Cournot setting where we assume that the tuple of strategies is unobservable while payoff functions and strategy sets are public knowledge through a common knowledge assumption. By utilizing observations of noise-corrupted prices, iterative fixed-point schemes are developed, allowing for simultaneously learning the equilibrium strategies and the misspecified parameter in an almost-sure sense.

math.OC↗

arXiv2017

Fastest Convergence for Q-learning

The Zap Q-learning algorithm introduced in this paper is an improvement of Watkins' original algorithm and recent competitors in several respects. It is a matrix-gain algorithm designed so that its asymptotic variance is optimal. Moreover, an ODE analysis suggests that the transient behavior is a close match to a deterministic Newton-Raphson implementation. This is made possible by a two time-scale update equation for the matrix gain sequence. The analysis suggests that the approach will lead to stable and efficient computation even for non-ideal parameterized settings. Numerical experiments confirm the quick convergence, even in such non-ideal cases. A secondary goal of this paper is tutorial. The first half of the paper contains a survey on reinforcement learning algorithms, with a focus on minimum variance algorithms.

eess.SY↗

arXiv2016

Error Estimates for the Kernel Gain Function Approximation in the Feedback Particle Filter

This paper is concerned with the analysis of the kernel-based algorithm for gain function approximation in the feedback particle filter. The exact gain function is the solution of a Poisson equation involving a probability-weighted Laplacian. The kernel-based method -- introduced in our prior work -- allows one to approximate this solution using {\em only} particles sampled from the probability distribution. This paper describes new representations and algorithms based on the kernel-based method. Theory surrounding the approximation is improved and a novel formula for the gain function approximation is derived. A procedure for carrying out error analysis of the approximation is introduced. Certain asymptotic estimates for bias and variance are derived for the general nonlinear non-Gaussian case. Comparison with the constant gain function approximation is provided. The results are illustrated with the aid of some numerical experiments.

math.NA↗

arXiv2016

Approximating a Diffusion by a Hidden Markov Model

For a wide class of continuous-time Markov processes, including all irreducible hypoelliptic diffusions evolving on an open, connected subset of $\RL^d$, the following are shown to be equivalent: (i) The process satisfies (a slightly weaker version of) the classical Donsker-Varadhan conditions; (ii) The transition semigroup of the process can be approximated by a finite-state hidden Markov model, in a strong sense in terms of an associated operator norm; (iii) The resolvent kernel of the process is `$v$-separable', that is, it can be approximated arbitrarily well in operator norm by finite-rank kernels. Under any (hence all) of the above conditions, the Markov process is shown to have a purely discrete spectrum on a naturally associated weighted $L_\infty$ space.

math.PR↗

arXiv2016

Differential TD Learning for Value Function Approximation

Value functions arise as a component of algorithms as well as performance metrics in statistics and engineering applications. Computation of the associated Bellman equations is numerically challenging in all but a few special cases. A popular approximation technique is known as Temporal Difference (TD) learning. The algorithm introduced in this paper is intended to resolve two well-known problems with this approach: In the discounted-cost setting, the variance of the algorithm diverges as the discount factor approaches unity. Second, for the average cost setting, unbiased algorithms exist only in special cases. It is shown that the gradient of any of these value functions admits a representation that lends itself to algorithm design. Based on this result, the new differential TD method is obtained for Markovian models on Euclidean space with smooth dynamics. Numerical examples show remarkable improvements in performance. In application to speed scaling, variance is reduced by two orders of magnitude.

eess.SY↗

arXiv2016

Rationally inattentive control of Markov processes

The article poses a general model for optimal control subject to information constraints, motivated in part by recent work of Sims and others on information-constrained decision-making by economic agents. In the average-cost optimal control framework, the general model introduced in this paper reduces to a variant of the linear-programming representation of the average-cost optimal control problem, subject to an additional mutual information constraint on the randomized stationary policy. The resulting optimization problem is convex and admits a decomposition based on the Bellman error, which is the object of study in approximate dynamic programming. The theory is illustrated through the example of information-constrained linear-quadratic-Gaussian (LQG) control problem. Some results on the infinite-horizon discounted-cost criterion are also presented.

math.OC↗