Cooperation Enforcement Leveraging Evolutionary Game Theory for Vehicular Networks

This work concerns the design, testing, and comparison of approaches promoting the store-and-forwarding of messages in a Vehicular Delay-Tolerant Network (VDTN) while still making the network robust against attacks coming from ill-intentioned nodes that try to send large volumes of selfgenerated messages to cause congestion. Performance evaluation consisted of simulations with varying percentages of flooding and intermittent flooding attackers using the real-world road map of Salamanca on the Opportunistic Network Environment (ONE) simulator. Different performance evaluation metrics were considered, namely Message delivery ratio (MDR), latency, hop count, false positives in attacker detection, false negatives, reputation, and trust threshold. Evaluation metrics showed that the two novel approaches, designed using a reputation enforcement system that considers the previously known actions of opponents to isolate nodes that do not send messages from others, and also using concepts of Evolutionary Game Theory (EGT) to make the network adaptable to different states of the population of nodes, exceed the performance of the classic Tit-for-tat (TFT) and Tit-for-two-tats (TF2T) approaches for scenarios with or without attackers.