Research On Two-layer Distributed Fusion Decision-making Method Based On Evidence Theory

With the social development and progress,more and more organizations pursue high efficiency,low cost,and human-based management,and traditional concentrated decision-making is less able to meet organizations' needs of scientific,systemic and domestic decision,on which ground,distributed decision-making comes into being and is becoming an important decision making method in the field of decision making science with its advantage on parallel processing of information and pluralistic integration.Today,studies on distributed decision-making have attracted more and more attention and also achieved some accomplishments,but they do not apply to the new distributed decision-making problems which increasingly occur in current organizations.Therefore,systemic and thorough study on new distributed decision-making method which applies to current organizations' decision problems is of theoretical and practical significance.Evidence theory developed upon traditional probability theory is able to flexibly measure "uncertain","unknown" judges out of people's cognition,fuses judges of different reliability by its relatively simple fusion mechanism and results in multi-source reasoning information fusion.Therefore,Evidence theory as an uncertain information reasoning technology is widely used in the field of information fusion and decision science,etc.This paper studies double layer distributed decision-making problems which have prominent hierarchy and ordinal implementation of decision,on the ground that double layer distributed organization structure is gradually flattened.Evidence theory is the basic theory in this paper,combined with knowledge of decision science,management science,science of human behavior in organization,science of system engineering,etc.After carefully reviewed and studied a large number of literature materials,the author refins the research topic and finds the breakthrough point.This paper firstly introduces the basics,formation rule,theory advantage of evidence theory and its relation with decision making,then analyzes the problem of decision reasoning and that of information redundancy in double layer distributed decision-making in present organizations,finally establishes distributed fusion decision-making fusion methods in simple and complicated decision situation according to the quantity of participated departments,complexity degree of hierarchy in actual decision-making.In the process of constructing distributed fusion decision-making methods,firstly,the question description is made and decision model is suggested according to specific features of the decision problem based on decision science.Then,reasonable assumptions of relations between decision makers are given,reasoning information description mechanism of flexible communication advantage is suggested based on evidence theory,and information fusion related theorems and corollaries are given.Finally,according to top-down decision order,procedures of double layer distributed fusion decision-making are construsted.Data analysis are made and the result proves that methods suggested in this paper are scientific and practical.