Abstract : Proposed originally as stochastic scheduling, the probabilistic real-time scheduling concerns real-time systems with at least one parameter described by a random variable. Any parameter of the task may have such description, but the existing work concentrates on the probabilistic worst-case execution times. This lecture will provide the main results for such systems in the case of one processor and a list of the open problems. The second part of the lecture deals with real-time systems that have stochastic description of the parameters. All the probabilistic operations are based on convolutions and the complexity of these operations may be a problem for realistic implementations. Sampling techniques providing decreased complexity are presented. The lecture ends with the presentation of the main open directions for probabilistic real-time systems and their impact on real-time systems in general.