Impact of Number of Artificial Ants in ACO on Network Convergence Time: A Survey

Samuel W Lusweti; Collins O Odoyo; Dorothy A Rambim

doi:10.34010/injiiscom.v3i1.8431

Samuel W Lusweti Department of Information Technology, School of Computing and Informatics, Masinde Muliro University of Science and Technology, Kenya
Collins O Odoyo Department of Information Technology, School of Computing and Informatics, Masinde Muliro University of Science and Technology, Kenya
Dorothy A Rambim Department of Information Technology, School of Computing and Informatics, Masinde Muliro University of Science and Technology, Kenya

DOI: https://doi.org/10.34010/injiiscom.v3i1.8431

Keywords: Convergence time, Ant Colony Optimization, artificial ants, networks, parameter

Abstract

Due to the dynamic nature of computer networks today,
there is need to make the networks self-organized. Selforganization can be achieved by applying intelligent
systems in the networks to improve convergence time.
Bio-inspired algorithms that imitate real ant foraging
behaviour of natural ants have been seen to be more
successful when applied to computer networks to make
the networks self-organized. In this paper, we studied
how Ant Colony Optimization (ACO) has been applied
in the networks as a bio-inspired algorithm and its
challenges. We identified the number of ants as a
drawback to guide this research. We retrieved a number
of studies carried out on the influence of ant density on
optimum deviation, number of iterations and
optimization time. We found that even though some
researches pointed out that the numbers of ants had no
effect on algorithm performance, many others showed
that indeed the number of ants which is a parameter to
be set on the algorithm significantly affect its
performance. To help bridge the gap on whether or not
the number of ants were significant, we gave our
recommendations based on the results from various
studies in the conclusion section of this paper