An Adjustable Frequency Bat Algorithm Based on Flight Direction to Improve Solution Accuracy for Optimization Problems

Abstract

An Adjustable Frequency Bat Algorithm (AFBA) is proposed to improve solution accuracy for optimization problem in this study. The conception is to employ the adjustable frequency determined by flight direction of bats to adapt the velocity toward the correct direction. The bats emit an ultrasound with various frequencies decided by flight direction to the current best bat. The adjustable frequency can provide the bats correct direction, proper velocity to move their position. And the bats can more systematical explore new possible better position in movement. Subsequently, there are many scenarios designed by different dimensions from low to high and benchmark functions with diverse modal to verify the performance of the proposed AFBA. The experimental numeric result shows that AFBA has better ability of search to improve the quality of the global optimal solution than BA. The fitness errors almost are less than 1.00E-6 for the unimodal function and multimodal function in tested dimensions.