Non-linear Optimization on a Parallel Intel i860 RISC Based Architecture

Authors:
J. F. Ball;R. E. Dorsey;J. D. Johnson
Affiliations:
Mississippi Center for Supercomputing Research/;Department of Economics and Finance, University of Mississippi, Conner Hall, University, MS 38677, U.S.A./ e-mail: dorsey@bus.olemiss.edu/;Department of Management and Marketing
Venue:
Computational Economics - Special issue on computational economics in Geneva: volume 1: computational econometrics, statistics, and optimization
Year:
1997

Citing 3
Cited 0

Transputers and parallel architectures: message-passing distributed systems

Transputers and parallel architectures: message-passing distributed systems
Practical parallel computing

Practical parallel computing
The unpredictability of standard back propagation neural networks in classification applications

Management Science

Quantified Score

Hi-index	0.00

Visualization

Abstract

A description of a parallel implementation of the Genetic Algorithm fornon-linear optimization is presented. The system is implemented with fourIntel i860 RISC processors using INMOS transputers for communications. Theperformance is compared to the Cray Y-MP, Cray J916, an SGI Power ChallengeL R8000 75 MHZ workstation, an SGI Challenge L R4400 100 MHZ work-station, an SGI Indy SC R4400 200 MHZ workstation, and both parastation and Paramid parallel I860 machines, each with one to four participating I860s.Comparisons were made for neural network optimization of a limited dependentvariable problem and a chaotic time series. A complex rational expectationsmodel was also estimated. The I860 based machines were able to achieveperformances comparable to the supercomputers.