There has been recently a growing interest in probabilistic models for systems involoving concurrency. Such models are useful in various application areas, such as model checking and analysis of randomised protocols in general.

Among the various models of probabilistic Petri nets, or event structures, found in the literature, very few jointly consider probability and concurrency. In particular, the randomisation of runs modulo interleaving has been studied little (Volzer, Varacca, Benveniste). This is the basis of Markov nets.

The talk shows that geometric properties of occurrence nets and event structures have probabilistic counterparts, with full meaning w.r.t. concurrency. In particular, parallel processes are independent in the probabilistic sense. We will describe the randomisation and discuss some applications.