Effectiveness Measurement Of Three-Way Decision Based On Markov Model

Recently,the decision-making process of complex problems is fundamentally different from the past in terms of complexity,frequency,and breadth.At the same time,the uncertainty of decision-making problems is also increasing,and people’s traditional decision-making thinking and models have been impacted.The empirical decision-making method is challenging to meet the requirements of the big data era,and data-driven intelligent decision-making has become the central theme.Intelligent decision-making will benefit numerous fields such as public administration,transportation,and industrial manufacturing.How to transform the data advantage in big data into decision advantage is the current research hotspot of intelligent decision-making.It is difficult to efficiently deal with complex and uncertain problems by applying traditional methods such as machine learning and mathematical statistics to decision-making.However,the three-way decision theories and methods can effectively deal with complex and uncertain problems and are increasingly accepted by many scholars:attention and exploration.The primary thinking method of the three-way decision is"thinking with three."With the development of three-way decision,a three-way decision model of trisection,action,and outcome is gradually formed.Trisection means the whole is divided into three,and action implies the application of strategy,and the outcome is used to evaluate the impact of trisection and action.The general idea of the current research on the effectiveness measurement of the three-way decision is as follows:after applying the strategy,the difference between the benefits before and after the application of the policy is calculated,and the different measures the effectiveness.However,there is a cost to the imposition of strategies,leading to negative effects.In order to reduce the occurrence of this ineffective strategy and improve the effectiveness of the strategy application,the effectiveness of the strategy can be predicted before the strategy is applied.In response to this problem,this topic proposes a method for predicting the effectiveness of strategies based on the Movement-based three-way decisions（M-3WD）.（1）The process of predicting the effectiveness of the strategy is not affected by the subjective preference of the decision-maker,that is,the three-way decision model without object preference.This topic proposes an effective prediction method based on the Markov model.Relying on the TAO model,based on the M-3WD model,the measurement of effectiveness is carried out from a new perspective and perspective,and a new measurement framework is formed-a dynamic regional movement model with movement preferences.An effectiveness prediction method is proposed based on the three-way decision model with movement preference,an effectiveness prediction method is proposed.Firstly,aiming at the state transition of the three-partition in the moving process,an index to characterize the movement stability of the object-Movement effectiveness index（MEI）is proposed.Secondly,for each stage of dynamic movement,the Markov model is used to predict the state matrix of the three-partition.The three-partition benefit after the move is estimated by combining the movement benefit index.Based on this,it is decided whether to carry out this strategy so that the Performance metrics are converted to performance predictions before the move.Finally,specific examples are given to explain the effectiveness of the proposed model.（2）The process of predicting the effectiveness of the strategy is affected by the subjective preference of the decision-makers,namely the three-way decision model with object preference.This topic proposes an effective prediction method based on the fuzzy Markov model.The effectiveness prediction and analysis imposed by the policy is modeled using fuzzy Markov,through the correspondence between the three partitions of the movement and the states in the Markov model.Specifically,we propose a fuzzy set represented by three equal parts to express object satisfaction,denoted as U₊^（k）,U_～^（k）,U_-^（k）,respectively,corresponding to the object’s attitude as satisfied,uncertain,and dissatisfied.In addition,the membership relationship between the three partitions and the satisfaction set after each strategy is applied defined as the Strategy Effectiveness Index（SEI）.According to the fuzzy Markov model,the validity of the prediction is achieved,thereby avoiding the execution of ineffective and inefficient strategies.Finally,an example of a voting model is used to illustrate the feasibility of the proposed effectiveness measure.