Adaptive Selection of Cryptographic Protocols in Wireless Sensor Networks using Evolutionary Game Theory

Game theory applies to scenarios wherein multiple players with contrary motives contend with each other. Various solutions based on Game theory have been recently proposed which dealt with security aspects of wireless sensor networks (WSNs). However, the nodes have limited capability of rationality and evolutionary learning which makes it unfavorable to apply conventional game theory in WSNs. Evolutionary Game Theory (EGT) relies on bounded rationality assumption which is in harmony with the wireless sensor networks characteristics. Based on EGT, authors propose an adaptive security model for WSNs for the selection of cryptographic protocols during runtime. The authors formulate this selection in WSNs with the help of an evolutionary game to obtain the evolutionarily stable strategy (ESS) for the system. In this model, the sensor nodes dynamically adapt their defensive strategies to attain the most efficient defense, corresponding to the attackers varied strategies. Further, the simulations convey that the proposed system converges rapidly to the Evolutionary Stable Strategy. Not only the system converges, but also forms a stable system which was verified by deliberately destabilizing the system. Results show that the nodes quickly return to ESS even after perturbation.