HyperProb: A Model Checker for Probabilistic Hyperproperties

We present HyperProb, a model checker to verify proba- bilistic hyperproperties on Markov Decision Processes (MDP). Our tool receives as input an MDP expressed as a PRISM model and a formula in Hyper Probabilistic Computational Tree Logic (HyperPCTL). By restricting the domain of scheduler quantification to memoryless non-probabilistic schedulers, our tool exploits an SMT-based encoding to model check probabilistic hyperproperties in HyperPCTL. Furthermore, when the property is satisfied, the tool can provide a witness that can be used for synthesizing a DTMC that conforms with the specification.