Color matching of bicomponent (PET/PTT) filaments with high performances using genetic algorithm.

In recent years, a great interest has focused on the use of bicomponent filaments in several high-performance textile articles such as swimwear, sportswear and even high-quality denim. To dye fabrics containing these filaments, it is necessary to establish appropriate dye recipes allowing to obtain desired shades. In this article, we developed a genetic algorithm to optimize the color matching step of these bicomponent filaments, especially (PET/PTT) filaments. Three disperse dyes with different molecular weights were used for dyeing. The objective is to reproduce the reference color by choosing the appropriate disperse dyes among the available dyestuffs and their corresponding quantities to use on the mixture. For modeling, two sets of parameters (lied to the color formulation problem and the genetic algorithm), the objective function as well as the different stages of the algorithm were defined and described. In addition, different techniques of selection and mutation were applied and evaluated. The optimization criterion is to reduce the CMC color difference between the desired reference colors and the colors proposed by the algorithm. The developed algorithm showed good performances with very small color differences. The results indicate that the roulette wheel selection technique outperforms both rank and uniform selection methods. Moreover, employing a simple mutation strategy yields favorable outcomes with CMC color differences all lower than 1.