A balanced butterfly optimization algorithm for numerical optimization and feature selection

Butterfly optimization algorithm (BOA) is a relatively novel technique for solving function as well real-world applications. However, the paramount challenge in BOA that it prone to stagnation local optima. The purpose of this study balance exploration and exploitation abilities when two new strategies are introduced. dynamic inertia weight based on Logistic model first strategy introduced modify position updating equation. Another opposition-based learning. A variant called BBOA these proposed. Ten widely used benchmark test functions 30 complex benchmarks from CEC 2014 selected verify effectiveness BBOA. problems composed unimodal, multimodal, rotated, shifted, hybrid composite functions. experimental results analysis show proposed has better ability than conventional different characteristics Finally, applied solve three engineering applications sixteen feature selection problems. demonstrate can outperform other competitors terms accuracy solution constrained design