Parallel optimal statistical design method with response surface modelling using genetic algorithms

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)22_Publication in policy or professional journal

10 Scopus Citations
View graph of relations

Author(s)

  • A. Wu
  • K.Y. Wu
  • R.M.M. Chen
  • Y. Shen

Related Research Unit(s)

Detail(s)

Original languageEnglish
Pages (from-to)7-12
Journal / PublicationIEE Proceedings: Circuits, Devices and Systems
Volume145
Issue number1
Publication statusPublished - Feb 1998

Abstract

Genetic algorithms (GA) together with a boundary sampling strategy are proposed for optimal statistical design to achieve better performance and higher yield at minimum cost. Owing to the reduced number of circuit simulations, the proposed approach can provide a satisfactory model representation at improved computation speed for the selection of the response surface function approximation. Replacing circuit simulation with the proposed response function modelling method using GA, optimum statistical design is formulated as a problem that involves the solution procedures of design centring, fixed optimum tolerance assignment, and variable optimum-tolerance assignment. To achieve better computational efficiency a number of approaches for paralleling the genetic algorithm operations are identified and studied. The parallel GA is implemented on a parallel machine constructed from a cluster of networked workstations. An optimum statistical design example is presented to show the effectiveness of the proposed techniques. © IEE, 1997.

Research Area(s)

  • Boundary sampling, Genetic algorithm, Statistical design