"...A more rigorous way is
to define a spec, a succinct description of every allowable
behavior of the system, and prove that an implementation
always generates outputs consistent with the spec."

-IronFleet: Proving Practical Distributed Systems Correct

All of his papers are worth reading. He's looked into other data generation strategies beyond pure-random, as well as how to think about the effectiveness of the coverage of a PBT run.