Fast parallel iterative aggregation methods for simulation of dynamical systems

Wei K. Tsai, Garng Morton Huang, Lu Wei

Research output: Contribution to conferencePaper

Abstract

A novel iterative aggregation algorithm for the numerical simulation of dynamic systems is proposed and analyzed. The algorithm exploits the special structure of the linear equation problem resulting from the discretization of the dynamic system and of the aggregation/disaggregation procedures. The algorithm has a time complexity of (I(q) + 2M(q) + 3)log N in solving linear systems with q states for N discrete-time instants using O(qN) processors, where I(q) is the parallel time complexity for inverting a q × q matrix and M(q) is the parallel time complexity for matrix multiplication of two q × q matrices. The competing parallel cyclic reduction method for the same problem has a time complexity of I(q) + 3M(q) + 4)log N. Thus the proposed algorithm has a definite speed advantage.

Original languageEnglish
Pages2217-2222
Number of pages6
Publication statusPublished - 1 Dec 1989
Externally publishedYes
EventProceedings of the 1989 American Control Conference - Pittsburgh, PA, USA
Duration: 21 Jun 198923 Jun 1989

Other

OtherProceedings of the 1989 American Control Conference
CityPittsburgh, PA, USA
Period21/6/8923/6/89

Fingerprint

Dynamical systems
Agglomeration
Linear equations
Linear systems
Computer simulation

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Tsai, W. K., Huang, G. M., & Wei, L. (1989). Fast parallel iterative aggregation methods for simulation of dynamical systems. 2217-2222. Paper presented at Proceedings of the 1989 American Control Conference, Pittsburgh, PA, USA, .

Fast parallel iterative aggregation methods for simulation of dynamical systems. / Tsai, Wei K.; Huang, Garng Morton; Wei, Lu.

1989. 2217-2222 Paper presented at Proceedings of the 1989 American Control Conference, Pittsburgh, PA, USA, .

Research output: Contribution to conferencePaper

Tsai, WK, Huang, GM & Wei, L 1989, 'Fast parallel iterative aggregation methods for simulation of dynamical systems' Paper presented at Proceedings of the 1989 American Control Conference, Pittsburgh, PA, USA, 21/6/89 - 23/6/89, pp. 2217-2222.
Tsai WK, Huang GM, Wei L. Fast parallel iterative aggregation methods for simulation of dynamical systems. 1989. Paper presented at Proceedings of the 1989 American Control Conference, Pittsburgh, PA, USA, .
Tsai, Wei K. ; Huang, Garng Morton ; Wei, Lu. / Fast parallel iterative aggregation methods for simulation of dynamical systems. Paper presented at Proceedings of the 1989 American Control Conference, Pittsburgh, PA, USA, .6 p.
@conference{93b2eff8e6a74495b63106b13cd6d178,
title = "Fast parallel iterative aggregation methods for simulation of dynamical systems",
abstract = "A novel iterative aggregation algorithm for the numerical simulation of dynamic systems is proposed and analyzed. The algorithm exploits the special structure of the linear equation problem resulting from the discretization of the dynamic system and of the aggregation/disaggregation procedures. The algorithm has a time complexity of (I(q) + 2M(q) + 3)log N in solving linear systems with q states for N discrete-time instants using O(qN) processors, where I(q) is the parallel time complexity for inverting a q × q matrix and M(q) is the parallel time complexity for matrix multiplication of two q × q matrices. The competing parallel cyclic reduction method for the same problem has a time complexity of I(q) + 3M(q) + 4)log N. Thus the proposed algorithm has a definite speed advantage.",
author = "Tsai, {Wei K.} and Huang, {Garng Morton} and Lu Wei",
year = "1989",
month = "12",
day = "1",
language = "English",
pages = "2217--2222",
note = "Proceedings of the 1989 American Control Conference ; Conference date: 21-06-1989 Through 23-06-1989",

}

TY - CONF

T1 - Fast parallel iterative aggregation methods for simulation of dynamical systems

AU - Tsai, Wei K.

AU - Huang, Garng Morton

AU - Wei, Lu

PY - 1989/12/1

Y1 - 1989/12/1

N2 - A novel iterative aggregation algorithm for the numerical simulation of dynamic systems is proposed and analyzed. The algorithm exploits the special structure of the linear equation problem resulting from the discretization of the dynamic system and of the aggregation/disaggregation procedures. The algorithm has a time complexity of (I(q) + 2M(q) + 3)log N in solving linear systems with q states for N discrete-time instants using O(qN) processors, where I(q) is the parallel time complexity for inverting a q × q matrix and M(q) is the parallel time complexity for matrix multiplication of two q × q matrices. The competing parallel cyclic reduction method for the same problem has a time complexity of I(q) + 3M(q) + 4)log N. Thus the proposed algorithm has a definite speed advantage.

AB - A novel iterative aggregation algorithm for the numerical simulation of dynamic systems is proposed and analyzed. The algorithm exploits the special structure of the linear equation problem resulting from the discretization of the dynamic system and of the aggregation/disaggregation procedures. The algorithm has a time complexity of (I(q) + 2M(q) + 3)log N in solving linear systems with q states for N discrete-time instants using O(qN) processors, where I(q) is the parallel time complexity for inverting a q × q matrix and M(q) is the parallel time complexity for matrix multiplication of two q × q matrices. The competing parallel cyclic reduction method for the same problem has a time complexity of I(q) + 3M(q) + 4)log N. Thus the proposed algorithm has a definite speed advantage.

UR - http://www.scopus.com/inward/record.url?scp=0024926210&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0024926210&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0024926210

SP - 2217

EP - 2222

ER -