• Probability
• Petri Net
• Java language
• Confusion
• Concurrency

Concurrency and probability are two important features of modern ICT systems, models and languages.
The study of concurrent models where decisions are driven by probabilistic distribution is complicated by the so-called confusion problem. It arises when an event increases or decreases the set of alternatives for another concurrent event that is already enabled.
The confusion problem interferes with the concurrency, in the sense that, equivalent computations from point of view of concurrency, can be assigned different probabilities.
In a recent paper it has been shown that confusion can be removed by identifying the loci of decisions and introducing additional dependencies between events.
In the thesis we develop a tool for the automatic and compositional synthesis of confusion free models, exploiting the theory developed there.
The tool, developed in Java language, receives in input a Petri net (in a suitable class), removes automatically confusion, and produces in output the corresponding transformed Petri net.
The format file for exchanging Petri nets that we use is a standard XML dialect called PNML. Therefore, our tool reads a PNML representing Petri net to transform and outputs another PNML file describing the corresponding Petri net without confusion.
The output Petri net will be described also in DOT language, mainly for a better rendering of its graphical layout.