A stable bio-inspired resource assignment strategy for elastic traffic. (2nd February 2015)

• Record Type: Journal Article
• Authors: Goudarzi, Pejman
• Is Part Of: International journal of computational vision and robotics. Volume 5:Number 1(2015)
• Page Start: 55
• Page End: 71
• Publication Date: 2015-02-02
• Subjects: genetic algorithm -- evolutionary computation -- gravitational search algorithm -- GSA -- penalty function; Computer vision -- Periodicals; Robotics -- Periodicals; Artificial intelligence -- Periodicals; 006.3705
• Languages: English
• ISSNs: 1752-9131

Abstract

Nature-inspired methods have attracted many attentions today for solving difficult and non-linear engineering problems. Some important examples include gravitational search algorithm (GSA), genetic algorithm (GA), DNA computing, artificial immune systems, fuzzy logic etc. Fair network resource allocation strategies are based on solving a form of constrained optimisation problem. There are plenty of high-speed fair rate allocation methods in the literature, some of them are based on fuzzy controllers for improving the convergence speed and are not necessarily optimal. Hence, in the current research, some recent important heuristics such as GA, GSA and their fuzzy-based counterparts have been used for finding the optimal allocated rates. Stability analysis is presented to guarantee the convergence property of the algorithm. After simulating, the results show that using the mixed fuzzy-GSA approach improves the conventional methods in convergence speed and results in fewer oscillations.