Quantitative verification of beta reputation system using PRISM probabilistic model checker

Reputation systems are responsible for computing the reputation rank of entities in a community or network based on collecting the opinions. Reputation systems have gained lots of interests in different environments such as P2P networks and e-market-places. Despite the popularity of reputation systems, they are vulnerable to different kinds of attacks which can simply lead the system to erroneous results. In this paper we propose a novel approach for quantitative verification of reputation models using Prism probabilistic model checker. We have applied the proposed method to Beta reputation system as a famous and widely used reputation model that is the base of many other recent reputation models. The proposed method is capable of verifying the reputation model for finding the worst possible attack scenario. Also it can be used to find a series of pre-defined attacks. To illustrate the proposed method, three case studies are also presented.