Biomimetic adaptation of the evolutionary algorithm, NSGA-II-aJG, using the biogenetic law of embryology for intelligent optimization

Abstract

Several of the recent optimization techniques have been adapted from nature. The elitist nondominated sorting genetic algorithm with the adapted jumping gene operator (NSGA-II-aJG) is one such evolutionary technique inspired by genetics. This algorithm is quite useful for solving multiobjective optimization problems. The drawback of these techniques is the inordinately large amount of computational effort required for solving real-life problems, even though these techniques are quite robust as compared to conventional techniques. Their use for online optimization is particularly limited. Many industrial optimization problems require frequent changes in the objective functions as well as the decision variables, even though the system itself remains the same. Surprisingly, no algorithm has been developed which makes use of previous information for solving a different problem for the same system in a comparatively short (computational) time. The algorithm developed in this study, namely, B-NSGA-II-aJG, carries this out more intelligently using the biogenetic law of embryology. This algorithm increases the speed of convergence significantly.