Genetic Algorithms, Another Tool for Quad Mesh Optimization?

Proceedings, 7th International Meshing Roundtable, Sandia National Lab, pp.497-504, October 1998

MESHING
RESEARCH
CORNER

7th International Meshing Roundtable
October 26-28, 1998
Dearborn, Michigan, USA

Mike Holder, Ph.D. Student
Civil Engineering Dept., University of Alabama
Box 870205, Tuscaloosa AL 35407
mikehii@traveller.com

Jim Richardson, Assoc. Professor
Civil Engineering Dept., University of Alabama
jrichardson@coe.eng.ua.edu

Abstract
The purpose of this paper is to investigate the use of a genetic algorithm (GA) to perform the finite element analysis mesh smoothing process. It is the goal of this paper to take a simple quadrilateral mesh and smooth it using a GA into a useful model that will provide a correct solution. The GA smoothing technique is demonstrated on two simple quadrilateral mesh examples. Only one node is moved at a time in the examples.

A distortion metric is used to quantify the "goodness" of a quadrilateral element and serves as the fitness function for the GA. Other implementation details such as convergence criteria, population size, cross-over probability and mutation rate are discussed in the paper. Early results from the two simple check problems are presented. Finally, planned future work is outlined and possible GA smoothing applications are discussed.

Download Full Paper (Postscript Format)