Verifying Systems-of-Systems with Statistical Model Checking

International audienceVerifying the correctness of systems-of-systems (SoS) is a key challenge, largely because SoSs are evolutionarily developed by combining autonomous systems to fulfill a mission that could not be performed by one of the constituent systems alone. In the trade-off of correctness vs. scalability, model checking does not scale up to address the trustworthiness of SoSs, owing to the state explosion problem. A recent technique, however, has overcome this shortcoming: ‘Statistical Model Checking’ is based on sampling traces of the system-of-interest until adequate statistical evidence has been established