Enhancement of Petri nets identified from discrete-event processes by scheduling t-components

The enhancement of cyclic Petri nets (PN) identified from a set S of event sequences generated by discrete event processes is addressed. The refinement aims to yield a model that accurately describes the sequencing of job execution and represents a low surplus language. A novel method that extends discovery methods for concurrent repetitive processes dealing with long sequences of events that include all variants process behaviour, is presented. Given a PN N and its set of minimal t-invariants Y, the method analyses the execution of the t-components in every sequence in S; it determines patterns of representative transitions S Y that schedule the execution of t-components in N . Afterwards, S Yi are treated to identify a new PN N ’ that uses S Y and new places; N ’ schedules the occurrence of representative transitions in N . The enhanced model N ’’ is obtained by merging the representative transitions of both N and N ’.