Online parameter tuning using Particle Swarm Optimization for ant-based QoS routing in mobile ad-hoc networks

Abstract

Parameter tuning of metaheuristic is the process of finding and controlling correct combination and values of an algorithm's parameters for each individual problem. Since the performance of Ant Colony Optimization (ACO) is influenced by its parameter values, many techniques were proposed in the literature to tune the parameters in ACO. This is because parameters can implicitly determine the amplification and diversification of the search process. ACO is applied to a variety of optimization problems and, unfortunately, there are no universal parameter values which can be used in ACO to solve all kinds of real-world optimization problems efficiently and effectively due to the differences in size and type of these real-world applications. In this paper, we present a mechanism using Particle Swarm Optimization (PSO) to adaptively tune the parameters of ACO using different ranges for each parameter. The parameter-tuned ACO is applied to provide Quality of Service routing in mobile ad-hoc network (MANET). The performance of the parameter-tuned ACO is compared with a non-adaptive ACO version.