What is Cluster-Based Algorithms

Self-organization is a phenomenon in nature which has been studied in several areas, namely biology, thermodynamics, cybernetics, computing modeling, and economics. Systems exhibiting self-organization have well defined characteristics such as robustness, adaptability, and scalability, which make self-organization an attractive field of study for two kinds of applications: a) maintaining the communication among mobile devices in wireless networks, and b) coordination of swarms of mobile robots. In ad hoc networks, there is not necessarily an underlying infrastructure in which the nodes can maintain communicated with other nodes; so due to this feature, it is necessary to provide efficient self-organization algorithms for routing, managing, and reconfiguring the network. Furthermore, self-organization in nature provide clear examples about how complex behaviors can arise from only local interaction between entities, namely the ants colony, feather formation, and flock of birds. Based on the above mentioned examples, several algorithms have been proposed to accomplish robot formations using only local interactions. Due to resource constraints in mobile devices, selforganization requires simple algorithms for maintaining and adapting wireless networks. The use of resources for establishing robot formations can be reduced by improving simple rules to accomplish the formation. This article first presents a brief overview of several works developed in ad hoc networks; then, delves deeper into the key algorithms; and finally, challenges arising in this area are discussed.