Research On Dynamic Multi-criteria Decision-making Method Based On Fuzzy Cognitive Map

Driven by information technology,decision-making activities from many fields such as medical decision-making,financial risk assessment and environmental decision-making are essentially different from traditional decision-making activities in terms of frequency,breadth and complexity.It is urgent to solve challenges such as the impact of increasing uncertainties and the assessment of future evolution trends.Concerning the abovementioned challenges,this article develops a dynamic multi-criteria decision-making method based on fuzzy cognitive maps from the perspective of multi-discipline integration of multi-attribute decision-making and cognitive learning,evolutionary learning,and integrated learning.The specific research contents are as follows:(1)Research the fuzzy cognitive map method based on improved Jaya algorithm learning.As a kind of graph structure that connects the causal events,participation values,goals and trends in the fuzzy feedback dynamic system through the arcs of various concepts,the fuzzy cognitive map has good reasoning and prediction capabilities.One of the keys is how to quantify the relationship matrix of the fuzzy cognitive maps.At present,the quantitative learning algorithm still needs to solve the challenges of local optimization and the difficulty of obtaining prior knowledge.Therefore,this article proposes a fuzzy cognitive map method based on improved Jaya learning.First,construct the topological structure of the initial fuzzy cognitive map.Secondly,combined with the principle of simulated annealing to improve the Jaya algorithm based on data-driven characteristics.It is not necessary to obtain prior knowledge,but to optimize the fuzzy cognitive map learning weight matrix in an adaptive self-learning manner.Finally,the time series data prediction process is realized through the reasoning mechanism of the fuzzy cognitive map.Experimental results prove that this method effectively avoids the local convergence of the Jaya algorithm,and improves the accuracy of prediction compared with other methods based on evolutionary algorithm learning fuzzy cognitive maps.(2)Research the dynamic multi-criteria decision-making method based on fuzzy cognitive map and TOPSIS.Considering the increasing uncertainties in the decision-making process and the challenges of evaluating future evolution trends,a dynamic multi-criteria decision-making method based on fuzzy cognitive maps and TOPSIS is proposed.First,use historical data to learn fuzzy cognitive maps to predict future decision-making information and mine decision-making trends.Secondly,determine the weight vector of the decision-making criteria at each stage,and calculate the priority score of the stage plan by using TOPSIS method.Finally,based on the characteristics of the skew distribution curve,the time sequence weight vector in the cycle is determined,and the optimal decision in the cycle is achieved by aggregating the decision results of each stage.Experiments show that the proposed method has learning and predictive performance.Compared with typical dynamic multi-criteria decision-making methods,the decision results are highly consistent,which verifies its effectiveness in solving data-driven decision-making problems.