An Efficient UD-Based Algorithm for the Computation of Maximum Likelihood Sensitivity of Continuous-Discrete Systems

Dimitri Boiroux; Rune Juhl; Henrik Madsen; John Bagterp Jørgensen

An Efficient UD-Based Algorithm for the Computation of Maximum Likelihood Sensitivity of Continuous-Discrete Systems

Dimitri Boiroux, Rune Juhl, Henrik Madsen, John Bagterp Jørgensen

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

This paper addresses maximum likelihood parameter estimation of continuous-time nonlinear systems with discrete-time measurements. We derive an efficient algorithm for the computation of the log-likelihood function and its gradient, which can be used in gradient-based optimization algorithms. This algorithm uses UD decomposition of symmetric matrices and the array algorithm for covariance update and gradient computation. We test our algorithm on the Lotka-Volterra equations. Compared to the maximum likelihood estimation based on finite difference gradient computation, we get a significant speedup without compromising the numerical accuracy.

Original language	English
Title of host publication	Proceedings of the 55th IEEE Conference on Decision and Control (CDC 2016)
Publisher	IEEE
Publication date	2016
Pages	3048-3053
ISBN (Print)	978-1-5090-1836-9
Publication status	Published - 2016
Event	55th IEEE Conference on Decision and Control - Las Vegas, United States Duration: 12 Dec 2016 → 14 Dec 2016 Conference number: 55 http://cdc2016.ieeecss.org/index.php

Conference

Conference	55th IEEE Conference on Decision and Control
Number	55
Country/Territory	United States
City	Las Vegas
Period	12/12/2016 → 14/12/2016
Internet address	http://cdc2016.ieeecss.org/index.php

OpenUrl availability

Full text

Cite this

@inproceedings{5eee5f17371d4151aeeb1bc6ff5fc2c3,

title = "An Efficient UD-Based Algorithm for the Computation of Maximum Likelihood Sensitivity of Continuous-Discrete Systems",

abstract = "This paper addresses maximum likelihood parameter estimation of continuous-time nonlinear systems with discrete-time measurements. We derive an efficient algorithm for the computation of the log-likelihood function and its gradient, which can be used in gradient-based optimization algorithms. This algorithm uses UD decomposition of symmetric matrices and the array algorithm for covariance update and gradient computation. We test our algorithm on the Lotka-Volterra equations. Compared to the maximum likelihood estimation based on finite difference gradient computation, we get a significant speedup without compromising the numerical accuracy.",

author = "Dimitri Boiroux and Rune Juhl and Henrik Madsen and J{\o}rgensen, {John Bagterp}",

year = "2016",

language = "English",

isbn = "978-1-5090-1836-9",

pages = "3048--3053",

booktitle = "Proceedings of the 55th IEEE Conference on Decision and Control (CDC 2016)",

publisher = "IEEE",

address = "United States",

note = "55th IEEE Conference on Decision and Control , CDC ; Conference date: 12-12-2016 Through 14-12-2016",

url = "http://cdc2016.ieeecss.org/index.php",

}

An Efficient UD-Based Algorithm for the Computation of Maximum Likelihood Sensitivity of Continuous-Discrete Systems. / Boiroux, Dimitri; Juhl, Rune; Madsen, Henrik et al.
Proceedings of the 55th IEEE Conference on Decision and Control (CDC 2016). IEEE, 2016. p. 3048-3053.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - An Efficient UD-Based Algorithm for the Computation of Maximum Likelihood Sensitivity of Continuous-Discrete Systems

AU - Boiroux, Dimitri

AU - Juhl, Rune

AU - Madsen, Henrik

AU - Jørgensen, John Bagterp

N1 - Conference code: 55

PY - 2016

Y1 - 2016

N2 - This paper addresses maximum likelihood parameter estimation of continuous-time nonlinear systems with discrete-time measurements. We derive an efficient algorithm for the computation of the log-likelihood function and its gradient, which can be used in gradient-based optimization algorithms. This algorithm uses UD decomposition of symmetric matrices and the array algorithm for covariance update and gradient computation. We test our algorithm on the Lotka-Volterra equations. Compared to the maximum likelihood estimation based on finite difference gradient computation, we get a significant speedup without compromising the numerical accuracy.

AB - This paper addresses maximum likelihood parameter estimation of continuous-time nonlinear systems with discrete-time measurements. We derive an efficient algorithm for the computation of the log-likelihood function and its gradient, which can be used in gradient-based optimization algorithms. This algorithm uses UD decomposition of symmetric matrices and the array algorithm for covariance update and gradient computation. We test our algorithm on the Lotka-Volterra equations. Compared to the maximum likelihood estimation based on finite difference gradient computation, we get a significant speedup without compromising the numerical accuracy.

M3 - Article in proceedings

SN - 978-1-5090-1836-9

SP - 3048

EP - 3053

BT - Proceedings of the 55th IEEE Conference on Decision and Control (CDC 2016)

PB - IEEE

T2 - 55th IEEE Conference on Decision and Control

Y2 - 12 December 2016 through 14 December 2016

ER -

An Efficient UD-Based Algorithm for the Computation of Maximum Likelihood Sensitivity of Continuous-Discrete Systems

Abstract

Conference

OpenUrl availability

Fingerprint

Cite this