Multi-Metric Clustering in Mobile Ad-Hoc Networks using Firefly Optimization Algorithm

Abstract

Mobile Ad-Hoc Network (MANET) is a form of a wireless network where the devices are dynamic and able to configure themselves on the fly. In such an environment, for supporting energy efficient and robust communication, it is expected that cluster based routing is a predominant solution. The cluster based routing divide the entire network into varying size groups known as clusters. While grouping nodes into clusters, it is essential to choose efficient clusterheads(CHs) based on their potential. One among the good metrics is that each node's associativity which can configure the mobility of a node in relation to its neighbor. The proposed work also gives importance to residual battery level and degree(number of neighborhood nodes) of each node to select the CHs. Finding optimal clusters in case of dynamic MANET is a type of NP-Hard problem which can get a solution by using various evolutionary algorithms, swarm optimization techniques, etc. One of the recent methods, Firefly optimization algorithm is applied in the proposed work. This optimization method will try to find out stable CHs with multiple objectives which are used in framing the fitness function. The different objectives considered in the proposed work will lead to the benefit that the proposed clustering algorithm suitable for both static and dynamic environment and also for the environment with a different battery capacity of the nodes. Evaluating against the genetic algorithm (GA) and particle swarm optimization (PSO) techniques, the proposed firefly algorithm shows better performance under different simulation scenarios.