CONSTRUCTS FOR MODELING INFORMATION-SYSTEMS WITH PETRI NETS

This paper is concerned with elementary and high-level Petri nets (coloured Petri nets and predicate/transition nets) and their utility for conceptual modeling of information systems. Particular features of existing net classes are discussed and additional net constructs are proposed for that purpose. Adequate conceptual models, i.e., models which reflect an intended view of an information system without any strange or artifical elements are best obtained if the following features are available to the modeler: (i) places have finite capacities (bounded places), (ii) arc expressions denote sets of variable cardinality, (iii) exclusive and shared use of side-conditions can be specified, (iv) terms denoting the set of all entities of a current marking or a specified subset thereof can be used in transition expressions, (v) static properties can be specified in state-oriented terms and impose restrictions on the dynamic properties specified in transition-oriented terms. The paper presents examples of each one of these desirables features and discusses them against the existing high-level Petri net classes or shows how they can be provided without resorting to so-called extensions.