Multi-source knowledge fusion framework: assessing suppliers’ research and development of complex products

Purpose In the customized supply chain of complex products, there is multi-source heterogeneous information with differential structures about the suppliers’ R&D performance from several experts and resident representatives. From the perspective of information fusion, this study aims to design a group decision-making platform describing and aggregating the suppliers’ multi-source R&D data. The aggregation result can reveal the suppliers’ R&D performance by fusing several sorts of heterogeneous appraisal knowledge. Design/methodology/approach A novel analytical model, house of multi-source knowledge fusion (HMKF), is designed as the analytical framework. Expanded to a group decision-making environment, an HMKF network is developed to realize the group fusion on the individual expert’s heterogeneous appraisal knowledge. Furthermore, a programming model is proposed to determine the indicator and expert weights by minimizing the knowledge fusion error. Finally, the case of a commercial aircraft supply chain is conducted to validate the above research, in which five system suppliers’ R&D performance are evacuated through three experts’ appraisal data and three resident representatives’ ranking knowledge. Findings HMKF shows the suppliers’ R&D performance information about Multi-Criteria Decision Analysis, appraisal confidence, pairwise comparison and ranking sequence. Networked HMKFs generate the group knowledge fusion method to aggregate experts’ appraisal knowledge. The optimal indicator weights and expert weights can help control the fusion gap. The appraisal results can assist the main manufacturer in precisely diagnosing the suppliers’ classified R&D performance and implementing continuous improvement in the future. Originality/value This study proposes a novel information fusion platform to synthesize multi-source heterogeneous information in group decision-making. The programming model minimizing the fusion gap is explored to determine the optimal weights and present supplies’ R&D performance through fully leveraging information.