Mobile Intelligence

18.2 RELATED WORK

In this section, we discuss few well-reviewed and recently proposed trust models [6–16]. Liu et al. proposed a trust model [6], in which evidence is collected by monitoring the behavior of one-hop nodes and also through recommendations received from other nodes. They represent trust in discrete values and also explore various approaches for distributing recommendations to other nodes. Pirzada et al. [7] proposed similar approach for establishing trusted routes in dynamic source routing (DSR) protocol [17]. However, they represent trust in continuous values. They assign trust weight to each link and then apply shortest path algorithm based on the assigned weights to choose a trusted route. Ant-based evidence distribution (ABED) [8] adopted swarm intelligence-based approach, which is modeled from ant colonies for distributing and discovering trust evidence. Mundinger and Boudec proposed a deviation test to handle honest-elicitation problem in Ref. [9], where the trust model at every node accepts recommendation only if it does not differ too much from their reputation ratings. To enforce reputation sharing through recommendations and to address honest-elicitation problem, Liu and Issarny [10] impose a node to rate recommendations based on its experience with the recommended node. They then encourage the node to share recommendations only with nodes that have been sharing honest recommendations. However, their proposal is not foolproof against colluding recommenders ...