Genetic algorithms with elitism-based immigrants for dynamic load balanced clustering problem in mobile ad hoc networks

Abstract

Abstract-Clustering can help aggregate the topology information and reduce the size of routing tables in a mobile ad hoc network (MANET). To achieve fairness and even energy consumption, each clusterhead should ideally support the same number of cluster members. Moreover, one of the most important characteristics in MANETs is the topology dynamics, that is, the network topology changes over time due to energy conservation or node mobility. Therefore, for a dynamic and complex system like MANET, an effective clustering algorithm should efficiently adapt to each topology change and produce the new load balanced solution quickly. The maintenance of the cluster structure should be as stable as possible to reduce overhead. It requires that the new solution should try to keep most of the good parts in the previous solution. In this paper, we propose to use elitism-based immigrants genetic algorithm (EIGA) to solve the dynamic load balanced clustering problem in MANETs. Each individual represents a feasible clustering structure and its fitness is evaluated based on the load balance metric. Immigrants are introduced to help the population to handle the topology dynamics and produce new and closely related solutions. The experimental results show that EIGA can quickly adapt to the environmental changes (i.e., the network topology change) and produce highquality solutions after each change.

State-of-the-Art Clustering Schemes in Mobile Ad Hoc Networks: Objectives, Challenges, and Future Directions

“…MANETs can be classified based on a mobility model such as group mobility [51], [56], [59], [60], regular mobility [46] and random mobility [25], [28], [29], [31], [34], [35], [39], [40], [42], [44], [45], [48], [50], [52], [53], [55], [58], [61]- [64]. The nature of events that take place in MANETs has a very strong impact on clustering process and the clustering process can be classified into proactive clustering [25], [26], [28], [29], [32], [43], [46], [48], [50], [52]- [55], [58]- [63], [65], reactive clustering [31], [36], [37], [40], [42], [47], [49], [51], [56], [57],…”

“…The nature of events that take place in MANETs has a very strong impact on clustering process and the clustering process can be classified into proactive clustering [25], [26], [28], [29], [32], [43], [46], [48], [50], [52]- [55], [58]- [63], [65], reactive clustering [31], [36], [37], [40], [42], [47], [49], [51], [56], [57], [66] and hybrid clustering [45]. Similarly the nodes of the network can be grouped based on the size of the cluster such as fixed [46], [59], balanced [30]- [32], [37], [57], unbalanced [55] and dynamic or random partition [29], [34]- [36], [39], [40], [42], [45], [53], [54], [64]. In cluster based routing, each cluster may have a CH so the cluster based routing can be classified based on availability of CHs [25], [26], ...…”

“…Similarly the nodes of the network can be grouped based on the size of the cluster such as fixed [46], [59], balanced [30]- [32], [37], [57], unbalanced [55] and dynamic or random partition [29], [34]- [36], [39], [40], [42], [45], [53], [54], [64]. In cluster based routing, each cluster may have a CH so the cluster based routing can be classified based on availability of CHs [25], [26], [28]- [30], [32]- [43], [45]- [57], [59]- [61], [64] and the CHs criteria not defined or having no CHs [27], [31], [44], [62], [63], [65], [66]. Clustering can be classified based on CH selection criteria such as node degree [29], [32], [40], [42], [43], [45], [46], [50], [51], [53], [55]- ...…”

“…In cluster based routing, each cluster may have a CH so the cluster based routing can be classified based on availability of CHs [25], [26], [28]- [30], [32]- [43], [45]- [57], [59]- [61], [64] and the CHs criteria not defined or having no CHs [27], [31], [44], [62], [63], [65], [66]. Clustering can be classified based on CH selection criteria such as node degree [29], [32], [40], [42], [43], [45], [46], [50], [51], [53], [55]- [57], [61], remaining energy [26], [32], [37], [43], [47], [49]- [54], [61], [64], trust value [53], [64], mobility [26], [34], [42], [43], [46], [47],…”

A MANET-based semantic traffic management framework for ubiquitous public safety networks

“…This is mostly due to the fact that the network performance dramatically degrades once the network becomes overloaded. In the existing MANET research, traffic load balancing and broadcast reduction are two widely employed techniques for the traffic management. The first technique focuses into distributing the network traffic among the nodes and avoid any ‘hotspot’ issues.…”