A Novel Comprehensive Similarity Measure for Networked Operation Sequences

With the increasing development of knowledge-based technologies in manufacturing, it has been shown as a promising way to solve problems by reusing related knowledge discovered from existing data. When it comes to computer aided process planning (CAPP), however, most efforts have been done on serial operation sequences, ignoring the urgent and practical requirement of networked operation sequences, which can provide more flexible process options. Therefore, we propose a novel comprehensive similarity measure for networked operation sequences in this paper. First, the information embedded in the networked operation sequences is deeply analyzed and classified into four types: operation similarity, precedence relationship, number of common operations, and structure similarity. For each kind of information, an intermediate similarity measure is present considering the specific information need of the networked operation sequence. Then, principal component analysis (PCA) is used to obtain a comprehensive metric without repeatedly calculating the common part of the aforementioned four types of information. Finally, the performance of this proposed measure is evaluated by a numerical illustration, in which a data set of different similarity cases is also built.