iHow Optimization Algorithm: A Human-Inspired Metaheuristic Approach for Complex Problem Solving and Feature Selection

Abstract

In this paper, we propose the iHow Optimization Algorithm (iHowOA), a novel metaheuristic algorithm inspired by human-like cognitive processes such as learning, knowledge acquisition, and experience-based decision-making. The iHowOA aims to enhance the exploration-exploitation balance inherent to solving complex optimization problems by mimicking how humans gather data, learn, and improve over time. We tested the algorithm on standard benchmark functions, including those from the CEC 2005 suite, to evaluate its performance in terms of convergence, computational efficiency, and solution accuracy. Furthermore, the Binary iHowOA (biHowOA) was employed for feature selection tasks, and its performance was compared with other popular optimization algorithms. The results show that iHowOA achieves superior performance, consistently finding optimal solutions while maintaining computational efficiency. The biHowOA also demonstrated strong capability in feature selection, providing reduced feature sets with minimal classification error. Our experiments confirm that iHowOA offers an effective solution for both continuous optimization and feature selectionchallenges.