Research On Aggregated Decision Making Models Of Dual-information Based On Bayesian Network

With the decision-making tasks more difficult and exchange of information more convenient,decision-making tends to brainstorming ways, and large group decision-making has become the focusof decision-making problems. In the process of the decision-making, there are uncertaindecision-making problems because of the complicacy of such problems and uncertainty of people. Inorder to solve these decision-making problems, this paper put forward a double information decisionmethod based on Bayesian network inference.In order to construct the structure of Bayesian network, this paper firstly proposed a newabsolute close correlation model based on the shortcomings of the existing relational model andanalyzed its calculation formula and properties, then a example have been gave to prove theeffectiveness and practicability of this new method. This method provides the basis to determine theweight of expert in constructing the structure of Bayesian network.Then a new method of correlation measures based on the absolute close correlation model havebeen provided, according to the expert knowledge to determine when according to the Bayesiannetwork structure experts to information is intuitionist fuzzy number forms of network structuregeneration problem, in order to determine the weight of each expert, based on the new absolute closecorrelation model defines a intuitionist fuzzy correlation measure formula, the nature of the newcomputation, the method for determination of the expert weight method to determine the rally expertinformation network structure, and gave the detailed steps of this method.Network generation later, for a class of uncertain multiple attribute decision making problems,namely expert decision information for some attribute value is not the only certain, but with a certainprobability distribution, at this time due to the complicacy of such problems and expert informationcompleteness, led professionals can’t give some attribute value of the probability distribution, in viewof this kind of problem, put forward the consideration of the double information based on bayesiannetwork inference decision method, first of all, according to the practical problems of bayesiannetwork diagram, reason out the unknown probability distribution of the attributes of the probabilitydistribution, and combining the expert decision-making information for the experts of the schemeevaluation value, through the establishment of the experts was of the scheme, the value of the modeland then get the comprehensive evaluation value of the final solution, solve the decision problem.Finally, a decision making model of aggregating information based on Bayesian network is putforward aiming at a kind of uncertain multi-attribute decision making problems. The interval of each comprehensive evaluated value is calculated according to the information given by the decision maker.Then the expression of comprehensive evaluated value is inferred by the information of the Bayesiannetwork. Finally, the comprehensive evaluated value of each alternative is calculated by theaggregated model of this paper.