Research On Decision-making Models And Decision-making Methods Based On Multigranulation Rough Sets

Nowadays,with the continuous progress and development of human science,especially the rapid development of information science and technology,it has become easier for people to access data.However,due to the different sources of data,these data have their own characteristics and representations and transition from a single structured data type to a complex structured data type.The Pawlak rough set model,which deals with uncertain data,is difficult to analyze massive and complex structure data.For this reason,scholars have proposed a series of extended rough set models for data with different bases and different structures.For the constructed model,it is necessary to propose an effective and rapid processing method for a large number of complex data,so as to discover useful knowledge in the complex data information and to help solve real-life practical problems based on the obtained knowledge.Rough set theory can only rely on the characteristics of the data set itself,and can approximately represent the data set with various uncertainties according to the division without any processing of the data.This feature has outstanding advantages in knowledge acquisition and decision analysis of uncertain data.In this paper,we explore and study the establishment of decision-making models and their corresponding decision-making methods based on several types of multi-granulation rough sets,respectively.This thesis includes the following three aspects :Firstly,to address the problem of double-quantitative decision analysis in multigranulation decision-theoretic rough set models,decision-theoretic rough set models are explored from a double-quantitative perspective in a multi-granulation approximation space.Multi-granulation double-quantitative decision-theoretic rough set models based on logistic operators are proposed as well as the basic properties and decision rules of the proposed models are investigated.The data set is used to compare and validate and illustrate the superiority of the proposed Multi-granulation double-quantitative decisiontheoretic rough set models in terms of classification performance.Secondly,to address the problem of double-quantitative decision analysis in multi-granulation decision-theoretic rough fuzzy set models,a probability-based multigranulation rough fuzzy set is constructed by introducing probability operators to fuzzy approximation objects in the multi-granulation framework.Furthermore,six multigranulation decision-theoretic rough fuzzy set models are proposed and the basic properties of each model and the corresponding three-way decision rules are studied.The relationship between the models is discussed and the different effects of the threshold parameters on the three decision rules of the proposed model are analyzed by example when the threshold parameters are in different situations.Finally,to address the problem of decision analysis in local multi-granulation covering decision-theoretic rough set models,a local covering decision-theoretic rough set model is constructed by introducing local rough set theory in the covering approximation space.Based on this model,four local multi-granulation covering decision-theoretic rough set models are proposed by introducing multi-granulation decision theory and exploring the important properties of each model and the relationship with existing rough set models.Furthermore,the impact on the decision region when the two decision threshold parameters are varied is studied through practical cases.Based on rough sets,this thesis discusses and studies the decision-making model and decision-making analysis in multi-granulation approximation space.In the multigranulation decision-theoretic rough set model,four new multi-granulation doublequantitative decision-theoretic rough set models are established;In the multi-granulation decision-theoretic rough fuzzy set model,six multi-granulation double-quantitative decision-theoretic rough fuzzy set models are proposed;In the local multi-granulation decision-theoretic rough set model,four local multi-granulation covering decisiontheoretic rough set models are constructed.The proposition of these models and the discussion of their corresponding decision analysis enrich the rough set theory to a certain extent,and further promote the application of rough set theory in the case of complex data.