GENERATING HIERARCHICAL MODELS BY IDENTIFYING STRUCTURAL SIMILARITIES

This chapter describes a decrease‐and‐conquer‐based method for installing hierarchy into an otherwise “flat” model. The lookup method identifies the structurally similar components of a model. The clustering method establishes hierarchy over a flat model by forking a module out of each identical component identified by the lookup method. Although a wide array of languages is available to model a software system, each model‐checking tool essentially supports only a specific modeling language. Consequently, the lookup and clustering methods described specifically target CPN models. These methods offer no additional advantage in using CPN models, and they can be adapted independently for any other language that defines a notion of hierarchy and structural similarity. The chapter introduces the problem and provides an insight into the solution. It introduces the basics of substitution transitions. Prior to proposing the lookup and clustering methods, related work is compiled. The various experimental results are plotted and discussed.