A Novel Metaheuristic: Jaguar Algorithm with Learning Behavior

The most powerful metaheuristics must be good at both exploitation and exploration. In the present day, metaheuristics are designed to reach a balance between these two capabilities for the sake of avoiding being trapped in the local optimum or unable to achieve convergence. For the first time in history, it is noteworthy that exploitation and exploration are both strong in the use of the Jaguar Algorithm (JA). In this paper, JA presents a simple but robust method inspired by the behaviors of jaguars. One feature of the jaguar is that once a jaguar is locked onto its prey, the jaguar moves directly and swiftly toward the target in the hunting area that has been established as its own territory. In addition, jaguars can hunt more efficiently when they take advantage of teamwork. This jaguar behavior parallels the behavior that makes JA more efficient than other well-known algorithms in exploiting and exploring. The experiment of this research reveals that an appropriate cooperation of jaguars could have various positive influences in regard to benchmark functions.