A Generalized Framework For Kullback-Leibler Markov Aggregation

Rana Ali Amjad; Clemens Blöchl; Bernhard Geiger

doi:https://doi.org/10.1109/TAC.2019.2945891

A Generalized Framework For Kullback-Leibler Markov Aggregation

Rana Ali Amjad, Clemens Blöchl, Bernhard Geiger^*

^*Korrespondierende/r Autor/in für diese Arbeit

Publikation: Beitrag in einer Fachzeitschrift › Artikel

Abstract

We propose an information-theoretic Markov aggregation framework that is motivated by two objectives: 1) The Markov chain observed through the aggregation mapping should be Markov. 2) The aggregated chain should retain the temporal dependence structure of the original chain. We analyze our parameterized cost function and show that it contains previous cost functions as special cases, which we critically assess. Our simple optimization heuristic for deterministic aggregations characterizes the optimization landscape for different parameter values.

Originalsprache	englisch
Aufsatznummer	8861026
Seiten (von - bis)	3068-3075
Seitenumfang	8
Fachzeitschrift	IEEE Transactions on Automatic Control
Jahrgang	65
Ausgabenummer	7
DOIs	https://doi.org/10.1109/TAC.2019.2945891
Publikationsstatus	Veröffentlicht - Jul 2020

ASJC Scopus subject areas

!!Electrical and Electronic Engineering
!!Control and Systems Engineering
!!Computer Science Applications

Zugriff auf Dokument

https://doi.org/10.1109/TAC.2019.2945891

http://www.scopus.com/inward/record.url?scp=85087520579&partnerID=8YFLogxK

Fingerprint Untersuchen Sie die Forschungsthemen von „A Generalized Framework For Kullback-Leibler Markov Aggregation“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren

Amjad, R. A., Blöchl, C., & Geiger, B. (2020). A Generalized Framework For Kullback-Leibler Markov Aggregation. IEEE Transactions on Automatic Control, 65(7), 3068-3075. [8861026]. https://doi.org/10.1109/TAC.2019.2945891

@article{ca1adc0173b747ccbbcea25606d8061e,

title = "A Generalized Framework For Kullback-Leibler Markov Aggregation",

abstract = "We propose an information-theoretic Markov aggregation framework that is motivated by two objectives: 1) The Markov chain observed through the aggregation mapping should be Markov. 2) The aggregated chain should retain the temporal dependence structure of the original chain. We analyze our parameterized cost function and show that it contains previous cost functions as special cases, which we critically assess. Our simple optimization heuristic for deterministic aggregations characterizes the optimization landscape for different parameter values.",

keywords = "Bisimulation, information theory, lumpability, Markov chain, model reduction, predictability",

author = "Amjad, {Rana Ali} and Clemens Bl{\"o}chl and Bernhard Geiger",

year = "2020",

month = "7",

doi = "https://doi.org/10.1109/TAC.2019.2945891",

language = "English",

volume = "65",

pages = "3068--3075",

journal = "IEEE Transactions on Automatic Control",

issn = "0018-9286",

publisher = "IEEE Control Systems Society",

number = "7",

}

TY - JOUR

T1 - A Generalized Framework For Kullback-Leibler Markov Aggregation

AU - Amjad, Rana Ali

AU - Blöchl, Clemens

AU - Geiger, Bernhard

PY - 2020/7

Y1 - 2020/7

N2 - We propose an information-theoretic Markov aggregation framework that is motivated by two objectives: 1) The Markov chain observed through the aggregation mapping should be Markov. 2) The aggregated chain should retain the temporal dependence structure of the original chain. We analyze our parameterized cost function and show that it contains previous cost functions as special cases, which we critically assess. Our simple optimization heuristic for deterministic aggregations characterizes the optimization landscape for different parameter values.

AB - We propose an information-theoretic Markov aggregation framework that is motivated by two objectives: 1) The Markov chain observed through the aggregation mapping should be Markov. 2) The aggregated chain should retain the temporal dependence structure of the original chain. We analyze our parameterized cost function and show that it contains previous cost functions as special cases, which we critically assess. Our simple optimization heuristic for deterministic aggregations characterizes the optimization landscape for different parameter values.

KW - Bisimulation

KW - information theory

KW - lumpability

KW - Markov chain

KW - model reduction

KW - predictability

UR - http://www.scopus.com/inward/record.url?scp=85087520579&partnerID=8YFLogxK

U2 - https://doi.org/10.1109/TAC.2019.2945891

DO - https://doi.org/10.1109/TAC.2019.2945891

M3 - Article

VL - 65

SP - 3068

EP - 3075

JO - IEEE Transactions on Automatic Control

JF - IEEE Transactions on Automatic Control

SN - 0018-9286

IS - 7

M1 - 8861026

ER -