Joint online parameter estimation and optimal sensor placement for the partially observed stochastic advection-diffusion equation

DOI10.1137/20M1375073zbMATH Open1484.35423arXiv2009.08693OpenAlexW3087503551MaRDI QIDQ5862897FDOQ5862897

Authors: Louis Sharrock, Nikolas Kantas 

Publication date: 10 March 2022

Published in: SIAM/ASA Journal on Uncertainty Quantification (Search for Journal in Brave)

Abstract: In this paper, we consider the problem of jointly performing online parameter estimation and optimal sensor placement for a partially observed infinite dimensional linear diffusion process. We present a novel solution to this problem in the form of a continuous-time, two-timescale stochastic gradient descent algorithm, which recursively seeks to maximise the log-likelihood with respect to the unknown model parameters, and to minimise the expected mean squared error of the hidden state estimate with respect to the sensor locations. We also provide extensive numerical results illustrating the performance of the proposed approach in the case that the hidden signal is governed by the two-dimensional stochastic advection-diffusion equation.

Full work available at URL: https://arxiv.org/abs/2009.08693

Recommendations

• Two-timescale stochastic gradient descent in continuous time with applications to joint online parameter estimation and optimal sensor placement
• scientific article; zbMATH DE number 4106758
• Optimal sensor placement for joint parameter and state estimation problems in large-scale dynamical systems with applications to thermo-mechanics
• Simultaneous state and parameter estimation and location of sensors for distributed systems
• A sparse control approach to optimal sensor placement in PDE-constrained parameter estimation problems

zbMATH Keywords

stochastic filteringonline parameter estimationoptimal sensor placementstochastic advection-diffusion equationtwo-timescale stochastic gradient descent

Mathematics Subject Classification ID

Computational methods for problems pertaining to probability theory (60-08) Stochastic partial differential equations (aspects of stochastic analysis) (60H15) Inference from stochastic processes and prediction (62M20) Signal detection and filtering (aspects of stochastic processes) (60G35) Reaction-diffusion equations (35K57) PDEs with randomness, stochastic partial differential equations (35R60) Identification in stochastic control theory (93E12) Optimal stochastic control (93E20)

Cites Work

• Adaptive subgradient methods for online learning and stochastic optimization
• Sequential Monte Carlo Methods in Practice
• Interpolation of spatial data. Some theory for kriging
• Statistics for spatio-temporal data
• ON STATIONARY PROCESSES IN THE PLANE
• An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach
• Inference in hidden Markov models.
• Title not available (Why is that?)
• Statistics for spatial data
• Particle approximations of the score and observed information matrix in state space models with application to parameter estimation
• Title not available (Why is that?)
• Optimal sensor networks scheduling in identification of distributed parameter systems
• Optimal Measurement Methods for Distributed Parameter System Identification
• Semigroups of linear operators and applications to partial differential equations
• On particle methods for parameter estimation in state-space models
• Optimal location of measurements for distributed parameter estimation
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Linear-Quadratic Optimal Actuator Location
• Title not available (Why is that?)
• Dynamics of evolutionary equations
• Sensors and controllers location in distributed systems - A survey
• Representation and control of infinite dimensional systems
• Infinite dimensional linear systems theory
• Estimation techniques for distributed parameter systems
• Stochastic Equations in Infinite Dimensions
• Ordinary Differential Equations with Applications
• Title not available (Why is that?)
• Fundamentals of stochastic filtering
• The Linear Regulator Problem for Parabolic Systems
• Title not available (Why is that?)
• Title not available (Why is that?)
• Stochastic approximation with two time scales
• A statistical modeling approach for air quality data based on physical dispersion processes and its application to ozone modeling
• Parameter estimation in general state-space models using particle methods
• Distributed parameter system indentification A survey†
• Title not available (Why is that?)
• Large-scale machine learning with stochastic gradient descent
• On minimum-contrast estimation for hilbert space-valued stochastic differential equations
• Title not available (Why is that?)
• An introduction to Morse theory. Transl. from the Japanese by Kiki Hudson and Masahico Saito
• Comparison of sequential data assimilation methods for the Kuramoto-Sivashinsky equation
• On the Poisson equation and diffusion approximation. III
• Partially observable linear systems under dependent noises
• Statistical inference for stochastic parabolic equations: a spectral approach
• Recursive identification in continuous-time stochastic processes
• Stochastic partial differential equations in groundwater hydrology. I: Theory
• Optimal sensor placement: a robust approach
• On-Line Parameter Estimation for Infinite-Dimensional Dynamical Systems
• Stochastic differential systems. I: Filtering and control. A function space approach
• The infinite-dimensional optimal filtering problem with mobile and stationary sensor networks
• Parameter identification in infinte dimensional linear systems
• Regularized Maximum Likelihood Estimate for an Infinite-Dimensional Parameter in Stochastic Parabolic Systems
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• A Survey of Infinite-Dimensional Filtering
• Optimal location of process measurements
• Title not available (Why is that?)
• Optimal sensor location problem for a linear distributed parameter system
• Title not available (Why is that?)
• Title not available (Why is that?)
• Observability and optimal measurement location in linear distributed parameter systems†
• Stochastic partial differential equations in groundwater hydrology. II: Application to Borden aquifer
• Statistical inference for SPDEs: an overview
• Title not available (Why is that?)
• Recursive algorithms for estimation of hidden Markov models and autoregressive models with Markov regime
• On the determination of optimal costly measurement strategies for linear stochastic systems
• Optimum measurements for estimation
• Approximation of the Algebraic Riccati Equation in the Hilbert Space of Hilbert–Schmidt Operators
• Estimation of Spatially Distributed Processes Using Mobile Spatially Distributed Sensor Network
• Parameter identification for hyperbolic stochastic systems
• Title not available (Why is that?)
• Infinite-dimensional filtering: The Kalman\3-Bucy filter in Hilbert space
• Title not available (Why is that?)
• Convergence of Galerkin approximations for operator Riccati equations - a nonlinear evolution equation approach
• Stochastic gradient descent in continuous time
• Maximum likelihood estimate for discontinuous parameter in stochastic hyperbolic systems
• Identification of an Infinite-Dimensional Parameter for Stochastic Diffusion Equations
• Galerkin Approximation for Optimal Linear Filtering of Infinite-Dimensional Linear Systems
• A stochastic advection-diffusion model for the Rocky Flats soil plutonium data
• Analyticity, Convergence, and Convergence Rate of Recursive Maximum-Likelihood Estimation in Hidden Markov Models
• Optimal control and observation locations for time-varying systems on a finite-time horizon
• Sensor Choice for Minimum Error Variance Estimation
• Maximum likelihood estimation in linear infinite dimensional models
• Online Maximum-Likelihood Estimation of the Parameters of Partially Observed Diffusion Processes
• Particle Filtering for Stochastic Navier--Stokes Signal Observed with Linear Additive Noise
• An introduction to stochastic Navier-Stokes equations
• Solutions and approximations to the Riccati integral equation with values in a space of compact operators
• Stochastic partial differential equation based modelling of large space-time data sets
• Recursive Maximum Likelihood Algorithm for Dependent Observations
• Score-Based Parameter Estimation for a Class of Continuous-Time State Space Models
• Asymptotic Properties of Recursive Particle Maximum Likelihood Estimation
• Title not available (Why is that?)
• Optimal sensor selection in sequential estimation problems

Cited In (6)

• Optimal sensor placement for joint parameter and state estimation problems in large-scale dynamical systems with applications to thermo-mechanics
• Sensor placement for optimal estimation of vector-valued diffusion processes
• Online parameter estimation for the McKean-Vlasov stochastic differential equation
• An extended PDE-based statistical spatio-temporal model that suppresses the Gibbs phenomenon
• Continuous‐time stochastic gradient descent for optimizing over the stationary distribution of stochastic differential equations
• Two-timescale stochastic gradient descent in continuous time with applications to joint online parameter estimation and optimal sensor placement

Uses Software

• AdaGrad
• ADADELTA

This page was built for publication: Joint online parameter estimation and optimal sensor placement for the partially observed stochastic advection-diffusion equation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5862897)