Maintaining and analyzing production process definitions using a tree-based similarity measure

Abstract

In this work a Case-Based reasoning system for managing production processes, declarative production process definitions in particular, with main focus on analysis and maintenance is introduced whereby each process task is represented by a case. A single process task definition includes among other elements, formulas, represented by fragmental program code. To get a meaningful similarity function among such cases, a new fuzzy tree edit distance metric on the formulas’ abstract syntax tree has been developed. The fuzzy tree edit distance addresses two aspects of similarity – similarity in terms of similar structure and similarity in terms of similar wording. As such, the proposed method represents a multidisciplinary approach to production process maintenance that includes methods from Case-Based reasoning and code clone detection.