选择合适的合作伙伴对提升供应链竞争力具有重要的战略意义。选择合作伙伴的常见方法有:线性加权法、数学规划法和层次分析法等,在存在噪声的实际决策环境中,大多数评价和选择方法缺乏对误分类的容错能力。本文根据三枝决策粗糙集理论,提出了一种新的方法来选择合作伙伴。该方法引入了最小风险贝叶斯决策理论,采用定量的概率包含关系来度量对象集合相对于目标概念的隶属度(条件概率),找出最小期望风险的决策,将其作为划分正域、负域和边界域的基本依据。本文给出了该方法的具体步骤,并通过算例对其有效性进行了验证。 Choosing the right partner is of strategic importance to enhancing the competitiveness of the supply chain. Common methods for selecting partners include linear weighting, mathematical programming and analytic hierarchy process. Most evaluation and selection methods lack the fault tolerance of misclassification in the actual decision-making environment with noise. Based on the theory of three branches decision-making rough set theory, this paper proposes a new method to select partners. This method introduces the least-risk Bayesian decision-making theory, uses the quantitative probability inclusion relation to measure the membership degree (conditional probability) of the object set relative to the target concept and finds the decision of the minimum expected risk as the partition of positive and negative The basic basis of the domain and the boundary domain. This paper gives the specific steps of this method, and validates its validity through examples.