Multi-attribute Decision Making And Time Series Forecast Based On Aggregation Operators

Aggregation operators as indispensable tools, which are used for informationfusion, have been disseminated throughout various fields including decision analysis,combination forecasting, military operations, and so on. In view of the existingproblems of multi-attribute group decision making and nonlinear time series forecast,we propose the correspongding solving methods, which are based on the aggregationoperators. The main work and innovations of the dissertation are summarized asfollows:(1) We extend the Quasi-OWA operator to the case in which the input argumentis a continuous valued interval and present the continuous Quasi-OWA (C-QOWA)operator, which generalizes a wide range of continuous operators such as the C-OWAoperator, the C-OWH operator and the C-GOWA operator. Then an orness measure toreflect the optimistic degree of the C-QOWA operator is proposed. Moreover, somedesirable properties of the C-QOWA operator associated with its orness measure areinvestigated. In addition, we apply the C-QOWA operator to the aggregation ofmultiple interval arguments and obtain the weighted C-QOWA operator, the orderedweighted C-QOWA (OWC-QOWA) operator, the combined C-QOWA (CC-QOWA)operator. We present an approach to multi-attribute group decision making based onthe CC-QOWA operator, when the decision information are continuous valuesinterval.(2) In order to aggregate triangular fuzzy numbers, the fuzzy Bonferroni mean(FBM) operator is developed and its some special cases are discussed. Based on this,the fuzzy weighted Bonferroni mean (FWBM) operator and combined fuzzy weightedBonferroni mean (C-FWBM) operator are proposed. Meanwhile, some desirableproperties of these operators are investigated. With respect to multi-criteria groupdecision making in which the decision making information is given by triangularfuzzy numbers, a new decision making method is proposed based on FWBM operatorand C-FWBM operator. The advantage of the proposed method is its capability tocapture the interrelationship between attributes. The results of an empirical analysisshow that the developed method is feasible.(3) In view of multi-attribute decision making in linguistic environment, weextend the Bonferroni mean operators to the situation where the inputs are linguisticarguments and proposed2-tuple linguistic Bonferroni averaging (2TLBA) operator, weighted2TLBA (W2TLBA) operator and combined W2TLBA (C-W2TLBA)operator. Moreover, some desirable properties and special cases of these aggregationoperators are investigated. Then a decision-making approach is presented based onW2TLBA operator and C-W2TLBA operator. Meanwhile, a numerical experiment isgiven to prove the feasibility of the developed approach.(4) In view of unstable and nonlinear time series, a dynamic forecasting model isproposed in this paper, which integrate wavelet decomposition, OWA operator basedSlip Discrete Difference Equation Prediction Model (OWA-SDDEPM) and Markovmethods. In this model, the original time series are decomposed into differentfrequency channels by multi-scale wavelet. Then we predict the wavelet coefficientsof the low-frequency approximation with the SDDEPM and predict the waveletcoefficients of high frequency details by Markov method, respectively. Therefore,forecasting value of the original time series is obtained by wavelet reconstruction ofthe low and high frequency foresting results. The model is applied to forecasting WTIweekly crude oil prices. The research result shows that the proposed forecastingmodel could not only forecast holistic fluctuation frequency of time series effectivelybut also characterize the details of the time series. The forecasting accuracy of thismodel is much higher than any other wavelet-based models.The above research results not only enrich the content of aggregation operatortheory, develop the corresponding new methods for multi-attribute group decisionmaking and time series forecasting, and provide more sufficient scientific evidencefor applying these methods to solve practical problems.