Optimal design of colour formulation prediction for cotton fabrics based on NSGA-II and TOPSIS

Abstract

The prediction of colour formulation is an important step in reproducing the target colour. At present, there are relatively few researches on multi-objective colour formulation problem, and the colour matching accuracy needs to be improved. In this research, a multi-objective evolutionary meta-heuristic method based on the Fast and Elitist Multi-objective Genetic Algorithm (NSGA-II) was proposed to predict the target colour recipes. The method used dye concentration as a variable and included three objective functions: (1) minimising the CMC (Colour Measurement Committee) colour difference between the formulation colour and the target colour, (2) minimising the metamerism index, and (3) minimising the cost of the formulation. The algorithm could obtain the Pareto optimal solution set after iteration. On this basis, the best combination of formulations was selected from the optimal solution set by combining the Expert Scoring Method (ESM), Entropy Weight Method (EWM) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The prediction effect of the model was evaluated by taking cotton fabrics and reactive dyes actually used in plant as examples. The results showed that 87.5% of the formulations met the CMC colour difference value of no more than 1, the metamerism index of 90.0% of the formulations did not exceed 1, and the cost of 92.5% of the formulations was reduced relative to the maximum extent in the Pareto optimal solution set. Further studies should be focused on removing duplicate individuals to give better diversity in the Pareto optimal solution set.