A cyber–physical system prospect theoretic game through a VANET lens

Abstract

In this paper, the problem of attack mitigation in an intelligent transportation network or vehicular network is considered as a game. The player’s perception of uncertainty and decision making is studied under a subjective prospect theoretic (PT) model and an objective expected utility theory (EUT) model. A game where each player chooses one of two strategies with certain probabilities is analysed. The case where subjective players bias their choices of the probabilities using the Prelec weighting function $w(.)$ is considered and compared with the EUT based decisions and the effect of the framing effect function $\nu (.)$ and $w(.)$. The corresponding Nash equilibria (NE) are derived and found through the replicator dynamic equation. Under the Prelec function, the results agree with the previously published results that the defender is biased more into defending the more important road side units. However, under both the $w(.)$ function and the framing effect, the players' behaviour does not depend on the loss penalty parameter, and the Prelec function dominates the framing effect. For small $\alpha$ values, the players make conservative decisions compared to higher $\alpha$ values regardless of the effect of the framing function. For high $\alpha$ values the players are more certain in their decisions than the EUT players.