CR-SMTC: Privacy preserving collusion-resistant multi-party trust computation

The ever-increasing use of trust and reputation models has posed new challenges in distributed environments. One of these challenges is the computation of trust while preserving privacy of feedback providers. This is because of the fact that some people may report a dishonest value due to social pressure or fear of the consequences. In this paper, we propose a privacy-preserving collusion-resistant multi-party trust computation scheme which uses data perturbation and homomorphic encryption to preserve the privacy of feedbacks. Our scheme is consisted of two protocols for private summation (S-protocol) and inner product (P-protocol). Our protocols are resistant to collusion of up to m+1 and m+2 agents, respectively, where m is a configurable parameter. In addition, their computational complexities are O(nm) and O(n(m+h)), respectively, where n is the number of agents and h is the homomorphic encryption algorithm complexity. We compare our protocols with related works and show its superiority in terms of collusion-resilience probability as well as complexity.