Control theory and experimental design in diffusion processes

Giles Hooker, Kevin Lin, Bruce Rogers

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

This paper considers the problem of designing time-dependent real-time control policies for controllable nonlinear diffusion processes, with the goal of obtaining maximally informative observations about parameters of interest. More precisely, we maximize the expected Fisher information for the parameter obtained over the duration of the experiment, conditional on observations made up to that time. We propose to accomplish this with a two-step strategy: when the full state vector of the diffusion process is observable continuously, we formulate this as an optimal control problem and apply numerical techniques from stochastic optimal control to solve it. When observations are incomplete, infrequent, or noisy, we propose using standard filtering techniques to first estimate the state of the system and then apply the optimal control policy using the posterior expectation of the state. We assess the effectiveness of these methods in three situations: a paradigmatic bistable model from statistical physics, a model of action potential generation in neurons, and a model of a simple ecological system.

Original languageEnglish (US)
Pages (from-to)234-264
Number of pages31
JournalSIAM-ASA Journal on Uncertainty Quantification
Volume3
Issue number1
DOIs
StatePublished - Jan 1 2015

Fingerprint

Experimental design
Control theory
Control Theory
Design of experiments
Diffusion Process
Control Policy
Stochastic Optimal Control
Fisher Information
Action Potential
Nonlinear Process
Nonlinear Diffusion
Statistical Physics
Real time control
Optimal Policy
Numerical Techniques
Neurons
Optimal Control Problem
Neuron
Optimal Control
Physics

Keywords

  • Controlled diffusions
  • Design of experiments
  • Neuron dynamics
  • Stochastic optimal control
  • Stochastic population dynamics

ASJC Scopus subject areas

  • Statistics and Probability
  • Modeling and Simulation
  • Statistics, Probability and Uncertainty
  • Discrete Mathematics and Combinatorics
  • Applied Mathematics

Cite this

Control theory and experimental design in diffusion processes. / Hooker, Giles; Lin, Kevin; Rogers, Bruce.

In: SIAM-ASA Journal on Uncertainty Quantification, Vol. 3, No. 1, 01.01.2015, p. 234-264.

Research output: Contribution to journalArticle

@article{cadf32925ccd4ab49ae23a28263565fb,
title = "Control theory and experimental design in diffusion processes",
abstract = "This paper considers the problem of designing time-dependent real-time control policies for controllable nonlinear diffusion processes, with the goal of obtaining maximally informative observations about parameters of interest. More precisely, we maximize the expected Fisher information for the parameter obtained over the duration of the experiment, conditional on observations made up to that time. We propose to accomplish this with a two-step strategy: when the full state vector of the diffusion process is observable continuously, we formulate this as an optimal control problem and apply numerical techniques from stochastic optimal control to solve it. When observations are incomplete, infrequent, or noisy, we propose using standard filtering techniques to first estimate the state of the system and then apply the optimal control policy using the posterior expectation of the state. We assess the effectiveness of these methods in three situations: a paradigmatic bistable model from statistical physics, a model of action potential generation in neurons, and a model of a simple ecological system.",
keywords = "Controlled diffusions, Design of experiments, Neuron dynamics, Stochastic optimal control, Stochastic population dynamics",
author = "Giles Hooker and Kevin Lin and Bruce Rogers",
year = "2015",
month = "1",
day = "1",
doi = "10.1137/140962280",
language = "English (US)",
volume = "3",
pages = "234--264",
journal = "SIAM-ASA Journal on Uncertainty Quantification",
issn = "2166-2525",
publisher = "Society for Industrial and Applied Mathematics Publications",
number = "1",

}

TY - JOUR

T1 - Control theory and experimental design in diffusion processes

AU - Hooker, Giles

AU - Lin, Kevin

AU - Rogers, Bruce

PY - 2015/1/1

Y1 - 2015/1/1

N2 - This paper considers the problem of designing time-dependent real-time control policies for controllable nonlinear diffusion processes, with the goal of obtaining maximally informative observations about parameters of interest. More precisely, we maximize the expected Fisher information for the parameter obtained over the duration of the experiment, conditional on observations made up to that time. We propose to accomplish this with a two-step strategy: when the full state vector of the diffusion process is observable continuously, we formulate this as an optimal control problem and apply numerical techniques from stochastic optimal control to solve it. When observations are incomplete, infrequent, or noisy, we propose using standard filtering techniques to first estimate the state of the system and then apply the optimal control policy using the posterior expectation of the state. We assess the effectiveness of these methods in three situations: a paradigmatic bistable model from statistical physics, a model of action potential generation in neurons, and a model of a simple ecological system.

AB - This paper considers the problem of designing time-dependent real-time control policies for controllable nonlinear diffusion processes, with the goal of obtaining maximally informative observations about parameters of interest. More precisely, we maximize the expected Fisher information for the parameter obtained over the duration of the experiment, conditional on observations made up to that time. We propose to accomplish this with a two-step strategy: when the full state vector of the diffusion process is observable continuously, we formulate this as an optimal control problem and apply numerical techniques from stochastic optimal control to solve it. When observations are incomplete, infrequent, or noisy, we propose using standard filtering techniques to first estimate the state of the system and then apply the optimal control policy using the posterior expectation of the state. We assess the effectiveness of these methods in three situations: a paradigmatic bistable model from statistical physics, a model of action potential generation in neurons, and a model of a simple ecological system.

KW - Controlled diffusions

KW - Design of experiments

KW - Neuron dynamics

KW - Stochastic optimal control

KW - Stochastic population dynamics

UR - http://www.scopus.com/inward/record.url?scp=85049800791&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049800791&partnerID=8YFLogxK

U2 - 10.1137/140962280

DO - 10.1137/140962280

M3 - Article

AN - SCOPUS:85049800791

VL - 3

SP - 234

EP - 264

JO - SIAM-ASA Journal on Uncertainty Quantification

JF - SIAM-ASA Journal on Uncertainty Quantification

SN - 2166-2525

IS - 1

ER -