Research On Decision Making Method Of Rough Programming Based On Similarity Relation

Decision method is a key research issue in management. Constructing different roughprogramming models and obtaining their solving methods are very important for forminga reasonable decision making. Uncertainties caused by objective conditions and humanfactors exist widely in practical problems. In order to represent practical problems as wellas possible and get better decision results, many researchers commit themselves to studyon the programming theory and method in uncertain environments. The researches ofstochastic programming and fuzzy programming are relatively mature. Compared with thesuccessful application of rough set theory in data mining, the researches on roughprogramming are at the starting stage.In this dissertation, for solving problem in rough environment, we first analyze theessential characteristics of rough programming systematically; give the general expressionof rough programming. And we construct a rough programming model based on synthesiseffect by distinguishing direct effect and indirect effect. Then we analyze the necessity andfeasibility of the research on the solution method of rough programming based onsimilarity relation. Then we construct a rough programming model based on the greatestcompatible classes and synthesis effect, and discuss its constructing procedures andstrategies. Combining with fuzzy set theory, we construct a rough programming modelbased on fuzzy similarity relation. Finally, we analyze the effectiveness and characteristicsof this model by a case-based example. The result indicate that this model has goodinterpretability, can integrate the decision preference into decision making process, and itextends and enriches the existing rough programming theories and methods.