Evolutionary optimization of hybrid indices for fashion design

The color harmony and style novelty of fashion are both considered for fashion design problems. The characteristics of the problem are firstly analyzed. It drew the conclusion that this kind of problems contains explicit and implicit objectives simultaneously and the two objectives are not conflicting as typical multi-objective optimization problems. Secondly, the explicit expression of color harmony is presented and genetic algorithm is adopted to perform the optimization. Thirdly, interactive genetic algorithm is applied to optimize the style novelty and to obtain the evaluation of implicit index. The results obtained by the two methods respectively are integrated and the two indices are co-evolved. Lastly, the interactive interface of human-machine is given and the results compared with the optimization using merely interactive genetic algorithm show the validity of the presented algorithm.