Semi - Markov based node behavior model for isolating selfish neighbors in MANETs

Reliable dissemination of data in mobile ad hoc network necessitates high degree of connectivity among the nodes present in that environment. The existence of selfish nodes could drastically affect the network connectivity. However, the impact of selfish nodes on network connectivity has been studied either through reputation factor computed based on past history or conditional probability manipulated based on both the past and present behavior of nodes. Hence, a need arises for formulating a mathematical model that makes its future routing decision only based on the present context of the node. In this paper, we propose a Semi-Markov based mathematical model, which isolates selfish nodes exponentially. In particular, we stochastically analyze the network resilience with the aid of transition probability matrix. The effectiveness of the proposed model for network survivability has been verified through quantitative analysis. The numerical results, thus obtained make it evident that the network survivability decreases when the number of selfish nodes increases in an ad hoc scenario.