A novel multi-attribute three-way decision model with three-parameter interval grey number decision-theoretic rough sets
Yu Qiao,Lirong Jian,Hechang Cai
DOI: https://doi.org/10.1108/k-11-2023-2331
IF: 2.352
2024-06-14
Kybernetes
Abstract:Purpose To overcome the limitations of traditional multi-attribute decision making (MADM) methods, which only provide deterministic rankings of decision objects, this paper proposes a novel multi-attribute 3WD model. This model presents three-parameter interval grey number decision-theoretic rough sets (TPIGNDTRSs), aiming to offer a reasoned interpretation of loss functions in grey environments and ensure objective assessment of conditional probabilities. Design/methodology/approach Firstly, the traditional equivalence relation is replaced with the probabilistic dominance relation (PDR), categorizing decision objects into two state sets in DTRS for more objective conditional probabilities. Secondly, as the three-parameter interval grey number (TPIGN) introduces the most probable value on the basis of the traditional two-parameter interval grey number, it provides a more comprehensive method for describing grey information. Consequently, integrating TPIGN into DTRS refines the interpretations of loss functions in grey environments. Finally, by utilizing two main sorting techniques, relative kernel and degree of accuracy ranking and possibility ranking, two types of 3WD rules with TPIGNDTRSs, are constructed. Findings This study has successfully developed and validated a new multi-attribute 3WD model. The model was tested in two distinct domains: evaluating innovation efficiency in high-tech enterprises and recommending movies in a practical case. The findings reveal that the model can effectively integrate relevant information of high-tech enterprises, provide the government with enterprise-level assessments, and gather consumer preferences to recommend the most suitable movies. Research limitations/implications This study treats the loss function as grey information in the 3WD model but overlooks the grey nature of evaluation values, limiting its applicability. Additionally, the model's reliance on subjective expert judgments and historical data to establish the loss function may affect its objectivity. The implications of this research are that the novel model overcomes traditional MADM limitations, enhancing decision-making quality and efficiency in complex and grey scenarios. The model's successful application in evaluating high-tech enterprises and recommending movies illustrates its dual value in both theory and practice. Originality/value Initially, the model proposed in this study is of significant importance for the development of the 3WD field, as it successfully addresses the challenges of uncertain loss functions and unknown conditional probabilities in grey information environments. Moreover, by integrating the 3WD model with MADM problems, it has broken through the bottlenecks of traditional MADM methods, offering new perspectives and strategies for solving MADM issues. Therefore, this research not only advances theoretical research but also provides powerful tools for practical applications.
computer science, cybernetics