Efficient and Robust Models for Trust Evaluation in Open, Dynamic, and Sociable Environments

Trust evaluation is widely used as an assistance for decision-making in sociable environments, such as social networks, sensor networks, web services, e-commerce, Crowdsourcing, Peer-to-Peer (P2P) networks, Internet of Things (IoT), and cloud computing, where entities cooperate to accomplish their objectives. However, due to the open, dynamic and decentralised properties of sociable environments, entities only have a local view of the environment, making it difficult to evaluate other’s trustworthiness based on their limited interaction experiences. In this situation, entities usually search for trusted paths between a source entity and a target entity based on neighbours, which incurs problems of being disturbed by unreliable trust paths and subjective neighbours. Besides, entities will seek information about target entities from third-party advisors, which incurs problems like being misled by subjective advisors and being deceived by dishonest advisors. Overall, there are two challenging problems for trust valuation in sociable environments: 1) uncertainties caused by decentralization and subjective neighbours/advisors; and 2) risks caused by dishonest advisors.