Multicriteria requirement ranking based on uncertain knowledge representation and reasoning

Abstract

The abundance of complex natural language descriptions related to requirements data presents major challenges for requirement analysis. Knowledge graphs (KGs), as the latest achievement in symbolic studies, are widely used in various fields due to their rich semantic expressive abilities. However, existing methods that mine resource description framework (RDF) triples in classical KGs cannot effectively capture the fuzzy semantic information in the product development system. To address this, we propose a multicriteria requirement ranking method based on uncertain knowledge representation and reasoning (KRR). First, we defined a model for representing uncertain knowledge to organize diverse data from multiple sources. This process was accomplished by constructing the requirement ontology, which is based on the function-behaviour-structure (FBS) model and requirements modelling-related documents. Next, a knowledge representation approach named the fuzzy requirement knowledge graph (FRKG) was devised by combining attribute confidence and predicate fuzziness. Then, knowledge reasoning rules were designed to enhance the edges in the FRKG, unveiling potential relationships between nodes. Utilizing the enriched FRKG (EFRKG), we proposed a multicriteria requirement ranking method based on grey relational analysis (GRA). To validate the effectiveness of the proposed approach, we conducted a case study involving unmanned aerial vehicles (UAVs). Furthermore, the semantic extension capability of FRKG was evaluated, and a comparison with traditional multicriteria requirement ranking methods was performed to demonstrate the efficiency of the proposed approach from both perspectives.