Adaptive control of routing protocol in mobile ad-hoc network using regression model

The performance of the Mobile Ad-Hoc NETwork are influenced by numerous factors which includes topological changes in network, network size variation, mobility of nodes, etc. Since the design of routing protocol is carried out by considering a particular environment, no existing routing protocol performs well for all kind of possible environments and requirements. In this paper we proposes an adaptive protocol selector for MANET which helps in selecting an optimum routing protocol by learning the current network context and user requirement. Regression models of various degrees have been designed for the performance metrics which depicts the functional relationship between the influencing factors and protocol performance. This paper includes the discussion on the design of adaptive protocol selection module, simulated results of Protocol Regression Model (PRM) and the simulated results of adaptive protocol selection. The PRM predicts the routing protocol performance for the current network context and the algorithm which helps in determining the appropriate protocol for a particular environment. Here DSDV, AODV and DSR protocols are chosen for the evaluation process.