Adaptive Krill Herd Algorithm for Global Numerical Optimization

Abstract

A recent bio-inspired optimization algorithm, that is, based on the Lagrangian and evolutionary behavior of krill individuals in nature is called the Krill Herd (KH) Algorithm. Randomization has a key role in both exploration and exploitation of a problem using KH algorithm. A new randomization technique termed adaptive technique is integrated with Krill Herd algorithm and tested on several global numerical functions. The KH uses Lagrangian movement which includes induced movement, random diffusion, and foraging motion, and therefore, it covers a vast area in the exploration phase. And then adding the powerful adaptive randomization technique potent the adaptive KH (AKH) algorithm to attain global optimal solution with faster convergence as well as less parameter dependency. The proposed AKH outperforms the standard KH in terms of both statistical results and best solution.