An Enhanced Bat Algorithm with Random Walk for Solving Continuous Optimization Problems

Abstract

Bat algorithm (BA) has been a successful algorithm for continuous optimization problems. However, its exploitation and exploration powers still leaves something to be desired. This work enhanced BA's exploitation power by introducing two new random walk processes that made its local search more thorough and enhanced it's exploration power by introducing inertia weight to intensify its global search near the end of the optimization process. We called this new algorithm an enhanced BA. The performance of enhanced BA on 15 widely accepted benchmark functions was compared with those of the original BA and genetic algorithm and found to be better than those achieved by the original BA and genetic algorithm on most of those benchmark functions. The directions of our future works are toward applying this enhanced BA to practical, real-world engineering problems and toward hybridizing BA with some other meta-heuristic algorithms.