Evolutionary algorithm for matching the weave and color of woven fabrics

Abstract

Creating the sample of an attractive woven fabric is a creative activity. Although there are computer programs applied to create a woven fabric sample, they cannot be helpful in the most important and creative part of the procedure. In order to provide help in this creative part of the procedure, an automated procedure was developed, which using an evolutionary algorithm creates an attractive sample. The basic, here described mechanism of creating a woven fabric, is an evolutionary algorithm. Different fabric samples are units which have to “survive” the iterative procedure in the “environment” which is an aesthetic experience of the user. Analogous with natural evolution the more the user likes the sample, the greater its adaptability to the environment and the higher the prospects for “survival”. On the basis of investigations of presented and analyzed results it can be concluded that the interactive evolutionary algorithm can be used in the procedure of matching the weave and color of the fabric on the computer. This is indicated by comparisons of different algorithm versions and their comparison with the test generator. It is observable that there is a progress in the arithmetic mean of adequacy and the best adequacy from the first to the last generation, meaning that the user increasingly likes the samples during the algorithm or that the algorithm creates samples that he likes aesthetically more on the basis of his selection.