Inproceedings,
Optimization Of a GA and Within the GA for a 2-Dimensional Layout Problem
Proceedings on the First International Conference on Evolutionary Computation and Its Applications, (1996)
Abstract
A GA's performance on a specific problem is related to many factors, such as genetic opera-
tors and corresponding parameter settings and the representation of the problem on the chromo-
some. Optimization of these factors to improve the speed and robustness of search is essential to
successful application of a GA. The work reported here uses a two-level GA system (DAGA2) to
solve a practical problem: the conceptual layout of machines on a factory floor (not the detailed
design), based on a matrix of positive and negative relationships of various strengths between
machines. This is a difficult, high-dimensionality problem. Performances of a traditional parallel
GA system and our DAGA2 system are compared. The later one is compared with and without
the use of several different representations for the problem at various times and in various subpop-
ulations, demonstrating the strong contribution which the use of multiple representations makes to
solution of the problem. The authors argue the necessity of optimizing many aspects of the GA in
order to obtain useful solutions on such difficult, real problems.
