Advance Bat Algorithm Inspired Algorithm for Fast Convergence and High Accuracy in Solving Numerical Optimisation Problems

This paper presents the Advanced Bat Algorithm (ABA), a novel enhancement to the standard Bat Algorithm (BA) designed to address persistent challenges in numerical optimisation, notably slow convergence and susceptibility to local minima in high-dimensional search spaces. ABA introduces three key innovations: physics-based inertia weight modeling, a boundary-aware position update using sine functions, and adaptive boundary control. These modifications collectively improve the balance between exploration and exploitation, ensuring robust performance across a variety of benchmark functions. Extensive computational experiments demonstrate that ABA achieves faster convergence and higher accuracy than standard BA and leading population-based algorithms such as Particle Swarm Optimisation (PSO), Genetic Algorithm (GA), and Harmony Search (HS). Comparative analysis with recent BA variants, including Improved Bat Algorithm (IBA) and Hybrid Self-Adaptive Bat Algorithm (HSABA), further confirms ABA’s superior performance, especially in high-dimensional and complex optimisation scenarios. The findings suggest that ABA is a promising tool for real-world optimisation tasks in 2026, with potential applications in AI model tuning, predictive maintenance, and smart manufacturing.