A robust ranking approach for target setting of system requirements based on preference decomposition

• Published in: Computers & industrial engineering Vol. 205; p. 111161
• Main Authors: Wang, Lilin, Tong, Shurong, Jin, Zesheng, Zhang, Xinwei
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.07.2025
• Subjects: Preference model; Stochastic multicriteria acceptability analysis; System requirements; Target setting; UTilités Additives method; UTilités Additives method; Target setting; Preference model; Stochastic multicriteria acceptability analysis; System requirements
• ISSN: 0360-8352
• DOI: 10.1016/j.cie.2025.111161

•A robust approach for target setting is proposed to reduce the cognitive burden.•Customer preference modeling is conducted based on the UTA-INT method and HFLTs.•The SMAA decomposition algorithm is combined with the CP function to set targets.•The approach is applied to set targets for unmanned aerial vehicle development. An essential aspect of system development is the target setting for system requirements (SRs). In the context of quality function deployment, constructing customer preference (CP) functions typically requires eliciting comprehensive information, including relative weights of customer requirements (CRs), influence relationships between CRs and SRs, correlations between SRs and marginal utility functions of different SRs. This process imposes a significant cognitive burden on both designers and customers. In addition, issues related to robustness and computational efficiency in the target setting process remain inadequately addressed. To tackle these challenges, a robust ranking approach based on preference decomposition is proposed for SR target setting. First, the CP function is modeled through the UTilité Additives with interactions method, which captures correlations between SRs to measure customer satisfaction. Second, indirect preference information is elicited through pairwise comparisons of reference alternatives and expressed in hesitant fuzzy linguistic terms, thereby reducing cognitive burden. Third, a series of linear programming models are developed to perform preference decomposition and determine parameter intervals of the CP function. Fourth, the decomposition algorithm of stochastic multicriteria acceptability analysis is integrated to robustly rank feasible SR combinations with computational efficiency, ultimately yielding optimal SR targets. The feasibility and effectiveness of the proposed approach are demonstrated through its application to target setting in unmanned aerial vehicle development.