A supervisory architecture and hybrid GA for the identifications of complex systems

Linyu Yang, John Yen, Athirathnam Rajesh, Ken D. Kihm

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Genetic Algorithms (GA's) have been demonstrated to be a promising search and optimization technique. However, there are two issues regarding applying genetic algorithms to complex system identifications. The first issue is the high computational cost due to their slow convergence. The second issue is its scalability to deal with high dimensional model identification problems. To alleviate the difficulties, we propose a two-layer supervisory model optimization architecture and hybrid GA algorithms. The upper supervisory layer guides the low level optimization algorithm so that the optimization space of the algorithm is gradually reduced. The lower layer uses simplex-GA approach to perform search and numerical optimization within the range defined by the upper layer. Simplex is added as an additional operator of traditional GA to speed up the convergence. We have applied the proposed approach to tomographic reconstruction and the modeling of central metabolism, the results are satisfactory.

Original languageEnglish (US)
Title of host publicationProceedings of the 1999 Congress on Evolutionary Computation, CEC 1999
PublisherIEEE Computer Society
Pages862-869
Number of pages8
Volume2
DOIs
StatePublished - 1999
Event1999 Congress on Evolutionary Computation, CEC 1999 - Washington, DC, United States
Duration: Jul 6 1999Jul 9 1999

Other

Other1999 Congress on Evolutionary Computation, CEC 1999
CountryUnited States
CityWashington, DC
Period7/6/997/9/99

Fingerprint

Hybrid Genetic Algorithm
Large scale systems
Complex Systems
Identification (control systems)
Genetic algorithms
Genetic Algorithm
Simplex Algorithm
Model Identification
Numerical Optimization
Identification Problem
System Identification
Optimization Model
Metabolism
Optimization Techniques
Computational Cost
Optimization Algorithm
Scalability
Speedup
High-dimensional
Mathematical operators

All Science Journal Classification (ASJC) codes

  • Computational Mathematics

Cite this

Yang, L., Yen, J., Rajesh, A., & Kihm, K. D. (1999). A supervisory architecture and hybrid GA for the identifications of complex systems. In Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999 (Vol. 2, pp. 862-869). [782513] IEEE Computer Society. https://doi.org/10.1109/CEC.1999.782513
Yang, Linyu ; Yen, John ; Rajesh, Athirathnam ; Kihm, Ken D. / A supervisory architecture and hybrid GA for the identifications of complex systems. Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999. Vol. 2 IEEE Computer Society, 1999. pp. 862-869
@inproceedings{dc87a471807d4fb29d0da3479a19bdc0,
title = "A supervisory architecture and hybrid GA for the identifications of complex systems",
abstract = "Genetic Algorithms (GA's) have been demonstrated to be a promising search and optimization technique. However, there are two issues regarding applying genetic algorithms to complex system identifications. The first issue is the high computational cost due to their slow convergence. The second issue is its scalability to deal with high dimensional model identification problems. To alleviate the difficulties, we propose a two-layer supervisory model optimization architecture and hybrid GA algorithms. The upper supervisory layer guides the low level optimization algorithm so that the optimization space of the algorithm is gradually reduced. The lower layer uses simplex-GA approach to perform search and numerical optimization within the range defined by the upper layer. Simplex is added as an additional operator of traditional GA to speed up the convergence. We have applied the proposed approach to tomographic reconstruction and the modeling of central metabolism, the results are satisfactory.",
author = "Linyu Yang and John Yen and Athirathnam Rajesh and Kihm, {Ken D.}",
year = "1999",
doi = "10.1109/CEC.1999.782513",
language = "English (US)",
volume = "2",
pages = "862--869",
booktitle = "Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999",
publisher = "IEEE Computer Society",
address = "United States",

}

Yang, L, Yen, J, Rajesh, A & Kihm, KD 1999, A supervisory architecture and hybrid GA for the identifications of complex systems. in Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999. vol. 2, 782513, IEEE Computer Society, pp. 862-869, 1999 Congress on Evolutionary Computation, CEC 1999, Washington, DC, United States, 7/6/99. https://doi.org/10.1109/CEC.1999.782513

A supervisory architecture and hybrid GA for the identifications of complex systems. / Yang, Linyu; Yen, John; Rajesh, Athirathnam; Kihm, Ken D.

Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999. Vol. 2 IEEE Computer Society, 1999. p. 862-869 782513.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - A supervisory architecture and hybrid GA for the identifications of complex systems

AU - Yang, Linyu

AU - Yen, John

AU - Rajesh, Athirathnam

AU - Kihm, Ken D.

PY - 1999

Y1 - 1999

N2 - Genetic Algorithms (GA's) have been demonstrated to be a promising search and optimization technique. However, there are two issues regarding applying genetic algorithms to complex system identifications. The first issue is the high computational cost due to their slow convergence. The second issue is its scalability to deal with high dimensional model identification problems. To alleviate the difficulties, we propose a two-layer supervisory model optimization architecture and hybrid GA algorithms. The upper supervisory layer guides the low level optimization algorithm so that the optimization space of the algorithm is gradually reduced. The lower layer uses simplex-GA approach to perform search and numerical optimization within the range defined by the upper layer. Simplex is added as an additional operator of traditional GA to speed up the convergence. We have applied the proposed approach to tomographic reconstruction and the modeling of central metabolism, the results are satisfactory.

AB - Genetic Algorithms (GA's) have been demonstrated to be a promising search and optimization technique. However, there are two issues regarding applying genetic algorithms to complex system identifications. The first issue is the high computational cost due to their slow convergence. The second issue is its scalability to deal with high dimensional model identification problems. To alleviate the difficulties, we propose a two-layer supervisory model optimization architecture and hybrid GA algorithms. The upper supervisory layer guides the low level optimization algorithm so that the optimization space of the algorithm is gradually reduced. The lower layer uses simplex-GA approach to perform search and numerical optimization within the range defined by the upper layer. Simplex is added as an additional operator of traditional GA to speed up the convergence. We have applied the proposed approach to tomographic reconstruction and the modeling of central metabolism, the results are satisfactory.

UR - http://www.scopus.com/inward/record.url?scp=33646475346&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646475346&partnerID=8YFLogxK

U2 - 10.1109/CEC.1999.782513

DO - 10.1109/CEC.1999.782513

M3 - Conference contribution

AN - SCOPUS:33646475346

VL - 2

SP - 862

EP - 869

BT - Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999

PB - IEEE Computer Society

ER -

Yang L, Yen J, Rajesh A, Kihm KD. A supervisory architecture and hybrid GA for the identifications of complex systems. In Proceedings of the 1999 Congress on Evolutionary Computation, CEC 1999. Vol. 2. IEEE Computer Society. 1999. p. 862-869. 782513 https://doi.org/10.1109/CEC.1999.782513