Title: Bat algorithm with Gaussian walk

Authors: Xingjuan Cai; Lei Wang; Qi Kang; Qidi Wu

Addresses: Department of Control Science and Engineering, Tongji University, Shanghai, 201804, China; Complex System and Computational Intelligence Laboratory, Taiyuan University of Science and Technology, 030024, China ' Department of Control Science and Engineering, Tongji University, Shanghai, 201804, China ' Department of Control Science and Engineering, Tongji University, Shanghai, 201804, China ' Department of Control Science and Engineering, Tongji University, Shanghai, 201804, China

Abstract: Bat algorithm is a novel branch of evolutionary computation. Although there are several research papers that focus on this new algorithm, however, few of them concerns the high-dimensional numerical problems. In this paper, a new variant called bat algorithm with Gaussian walk (BAGW) is proposed aiming to solve this problem. In this variant, a Gaussian walk is employed in the local turbulence instead of the original uniform walk to improve the local search capability. Furthermore, to keep the high exploitation pressure, the velocity update equation is also changed. Finally, to increase the population diversity, the frequency is dominated by each dimension in our modification, as well as it is depended on the different bat in the standard version. To test the performance of our variant, four famous un-constraint numerical benchmarks are employed, and test on different dimensional cases, simulation results show our modification is effective.

Keywords: bat algorithm; Gaussian walk; local search capability; velocity update equation; simulation; bio-inspired computation.

DOI: 10.1504/IJBIC.2014.062637

International Journal of Bio-Inspired Computation, 2014 Vol.6 No.3, pp.166 - 174

Received: 10 Feb 2014
Accepted: 27 Feb 2014

Published online: 10 Jun 2014 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article