Performance Comparisons of Socially Inspired Metaheuristic Algorithms on Unconstrained Global Optimization

Abstract

In recent years, many efficient metaheuristic algorithms have been proposed for complex, multimodal, high-dimensional, and nonlinear search and optimization problems. Physical, chemical, or biological laws and rules have been utilized as source of inspiration for these algorithms. Studies on social behaviors of humans in recent years have shown that social processes, concepts, rules, and events can be considered and modeled as novel efficient metaheuristic algorithm. These novel and interesting socially inspired algorithms have shown to be more effective and robust than existing classical and metaheuristic algorithms in a large number of applications. In this work, performance comparisons of social-based optimization algorithms, namely brainstorm optimization algorithm, cultural algorithm, duelist algorithm, imperialist competitive algorithm, and teaching learning based optimization Algorithms have been demonstrated within unconstrained global optimization problems for the first time. These algorithms are relatively interesting and popular, and many versions of them seem to be efficiently used within many different complex search and optimization problems.