### Abstract

Computer simulation models consisting of systems of differential equations, or other mathematical models, can present special problems to numerical optimization methods. Derivatives are often unavailable, function evaluations can be extremely expensive (e. g. , 1 h on an IBM 3090), and the numerical accuracy of each function value may depend on a complicated chain of calculations and so be impractical to prespecify. This last point makes it difficult to calibrate optimization routines that use finite-difference approximations for gradients. A strategy for comparing optimization techniques for these problems is presented, and several interesting findings for quasi-Newton methods, simplex search, and others are reviewed.

Original language | English (US) |
---|---|

Title of host publication | Winter Simulation Conference Proceedings |

Editors | Arne Thesen, Hank Grant, David W. Kelton, Madison Univ of Wisconsin-Madison |

Publisher | ACM |

Pages | 391-401 |

Number of pages | 11 |

ISBN (Print) | 0911801324 |

State | Published - Dec 1987 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Chemical Health and Safety
- Software
- Safety, Risk, Reliability and Quality
- Applied Mathematics
- Modeling and Simulation

### Cite this

*Winter Simulation Conference Proceedings*(pp. 391-401). ACM.

}

*Winter Simulation Conference Proceedings.*ACM, pp. 391-401.

**TESTING STRATEGIES FOR SIMULATION OPTIMIZATION.** / Barton, Russell Richard.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - TESTING STRATEGIES FOR SIMULATION OPTIMIZATION.

AU - Barton, Russell Richard

PY - 1987/12

Y1 - 1987/12

N2 - Computer simulation models consisting of systems of differential equations, or other mathematical models, can present special problems to numerical optimization methods. Derivatives are often unavailable, function evaluations can be extremely expensive (e. g. , 1 h on an IBM 3090), and the numerical accuracy of each function value may depend on a complicated chain of calculations and so be impractical to prespecify. This last point makes it difficult to calibrate optimization routines that use finite-difference approximations for gradients. A strategy for comparing optimization techniques for these problems is presented, and several interesting findings for quasi-Newton methods, simplex search, and others are reviewed.

AB - Computer simulation models consisting of systems of differential equations, or other mathematical models, can present special problems to numerical optimization methods. Derivatives are often unavailable, function evaluations can be extremely expensive (e. g. , 1 h on an IBM 3090), and the numerical accuracy of each function value may depend on a complicated chain of calculations and so be impractical to prespecify. This last point makes it difficult to calibrate optimization routines that use finite-difference approximations for gradients. A strategy for comparing optimization techniques for these problems is presented, and several interesting findings for quasi-Newton methods, simplex search, and others are reviewed.

UR - http://www.scopus.com/inward/record.url?scp=0023590833&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0023590833&partnerID=8YFLogxK

M3 - Conference contribution

SN - 0911801324

SP - 391

EP - 401

BT - Winter Simulation Conference Proceedings

A2 - Thesen, Arne

A2 - Grant, Hank

A2 - Kelton, David W.

A2 - Univ of Wisconsin-Madison, Madison

PB - ACM

ER -