An Improved Bat Algorithm Based on Second Flight and Random Disturbance

In order to overcome the shortcomings of conventional the basic bat algorithm (BA), such as easily trapping in local optima and lower search accuracy, this paper proposed the improvement strategy based on the second flight and the random disturbance. After each bat makes a random disturbance to its position.Using the adaptive weight learning the previous flight speed, so that bats can search in a good direction to avoid a bad direction. In order to ensure the diversity of the population, there is a punishment of the improved algorithm for the worst part to carry out a second flight search without learning last speed. This improves the global search ability of the improved algorithm.The algorithm tested on seven distinct types of benchmark functions. The results show that the improved strategy has a great improvement on the optimization accuracy and global search ability. The proposed algorithm has a better convergence rate and optimization accuracy.