Gray Wolf and Krill Herd optimizations: Performance analysis and comparison

Abstract

Herding behavior is defined as a group of animals of similar size that migrate in the same direction and hunt together. Gray wolves are usually seen in packs. Each gray wolf in the herd has a distinct duty and a distinct name that reflects the task. Krill swarms form the basis of ocean ecology. There are two reasons for the movement of the Krill herd. The first reason is that difficult for other organisms to prey on Krill living in herds. Another compelling factor is the way Krill form vast herds and effortlessly seize their prey. Gray Wolf Optimization (GWO) is inspired by gray wolf herding behavior, while Krill Herd Optimization (KHO) is based on krill herding. In this study, GWO and KHO algorithms are examined in detail and it is decided whether they had sufficient success. The fact that the GWO and KHO algorithms are swarm -based is accepted as a common feature of the two algorithms. However, compared with GWO and KHO analysis, as well as 23 single -mode, multimodal, and fixed -size multimodal benchmarking optimization tests. In another hand, the success of the algorithms has been demonstrated by running them on various dimensions ({10, 20, 30, 50, 100, 500}). Additionally, the performances of the GWO and KHO are compared with Tree Seed Algorithm (TSA), Particle Swarm Algorithm (PSO), Jaya algorithm, Arithmetic Optimization Algorithm (AOA), Evolutionary Mating Algorithm (EMA), Fire Hawk Optimizer (FHO), Honey Badger Algorithm (HBA) algorithms. Moreover, all of the analyses are obtained in detail, complete with statistical tests and figures. As a result, while GWO and KHO algorithms show superior success in different test problems with their own characteristics, they are at a competitive level with many old and newly proposed algorithms today. In order to determine the success of the GWO and KHO algorithms, not only the classical test functions but also two different benchmark test sets are used. These are the CEC-C06 2019 functions and the big data problem, which is a current problem today. The same algorithms are run for both problems and rank values are obtained according to the average results. In CEC-C06 2019 functions, KHO achieved good results, while in big data problems, GWO achieved good results. In this study, the success of the GWO and KHO algorithms are examined in detail in three different experimental sets and it sheds light on researchers who will study with GWO and KHO algorithms.