Routing Approach using Machine Learning in Mobile Ad-Hoc Networks

摘要

Mobile ad-Hoc networks are very popular and lot of research has been performed to study various characteristics of ad-hoc networks. Mobile ad-hoc network is a wireless network that does not contain any central control to monitor the network activities. Hence the nodes are bound to self organize themselves into small networks for the data communication. Since there is no central entity controlling activities in the network, nodes have to act as router and host both to manage the network. There are some challenges like dynamic topology, bandwidth, energy constraint etc. Many routing protocol have been developed to help the network routing activities. Protocols are divided in to Proactive, reactive and hybrid protocols. In this paper we are going to discuss a classification algorithm called CART algorithm from machine learning to predict the pattern or the decision a node will take serving as a individual rational entity. A nodes and other parameters in the network are taken as individual entity and based on entropy value it is decided which attribute is capable of being a root node and govern the network Routing decisions.