Multi-objective optimization of raw silk parameters using a combined support vector regression-genetic algorithm-desirability function approach

This work deals with the multi-objective optimization of raw silk qualities using the desirability function approach. Raw silk qualities are primarily governed by silk cocoon characteristics. In this paper, the complexity of simultaneous optimization of conflicting raw silk properties are resolved using a hybrid multi-objective optimization model where predictive power of support vector regression and optimization capability of genetic algorithm are employed with desirability function. The individual desirability of raw silk qualities is assessed from the four silk cocoon properties such as defective cocoon percentage, shell ratio, cocoon weight and cocoon volume. Raw silk parameters such as filament length, tenacity, renditta and reelability are combined together to express overall desirability. The proposed multi-objective optimization model can determine the optimum cocoon parameters essential to produce good-quality raw silk.