【摘要】 水雾化粉末质量预报系统具有原始数据不全的特点,是一个不完备的信息系统。本文依据此特点,结合变精度粗糙集方法,在相容关系的基础上引入集对分析方法,提出一种基于集对势容差关系的扩充粗糙集方法,并应用于该信息系统的建模中,通过对α的调节和控制,有效保证了制粉生产条件属性容差类划分的准确性和灵活性。最后,本文以该模型为基础,使用基于属性重要性的贪心算法讨论了制粉生产条件属性约简和知识获取的方法、步骤和决策规则。经实际生产验证了获取的预报知识的合理性和有效性,对制粉工艺和生产真正起到了优化作用。更多还原

【Abstract】 The forecast system of water atomized alloy powders’ quality is an incomplete information system because of its incomplete original data.In view of the incomplete system,based on the analysis of variable rough set and set pair analysis,a new kind of generalized rough set based on set pair situation tolerance relation was proposed in this paper,and was used in the system modeling. By regulating and controlling the parameter α to ensure veracity and flexibility when marking up the powders’ productive condition attributes.Based on the proposed model,a cupidity algorithm of attribute significance was introduced and used to realize the powders’ productive condition attributes reduction and acquire the relative knowledge.The proposed model and algorithm were applied in the production and it shows the acquired knowledge with high rationality and validity can truly be to optimize the production process.更多还原