Research On Assisted Decision Making Model Based On Uncertain Information Fusion

With the increasingly complex and changeable battlefield environment,the effective use of target detection information to construct combat-aided decision models such as target value assessment,target threat assessment,and combat plan evaluation has become a research hotspot.The complex and volatile nature of the modern battlefield leads to an environment of deep uncertainty in combat-aided decision-making.The deep uncertainty is manifested by the uncertainty of decision information,the complete unknown of decision weight information,the dynamic nature of the decision environment,and the variability of the evaluation group’s cognition.These problems make it difficult to effectively construct combat-aided decision models,making it challenging to carry out the above research on assisted decision models.In recent years,the development of uncertain information fusion techniques has pushed the research of various decision-making techniques to a new high point,showing potential applications in different decision-making fields.To this end,this paper takes the construction of the auxiliary decision model as the research objective,the depth uncertainty of the auxiliary decision information as the technical difficulty,and the fusion of uncertain information as the research means,and the research aims to break through the critical technologies of target value assessment,target threat assessment and combat plan assessment under uncertain environment,and then improve the auxiliary decision capability.The main research of this paper includes the following aspects.(1)Research on target value assessment with decision-oriented information in the form of interval intuitionistic fuzzy sets: A target value assessment model based on interval intuitionistic fuzzy set and evidential reasoning algorithm is proposed to address the problems of uncertainty of assessment results and complete unknown information of decision attribute weights in interval intuitionistic fuzzy environment leading to the failure of target value assessment.Firstly,the proposed model constructs a target value assessment system.Secondly,to solve the problem of entirely unknown attribute weight information,a target attribute weight determination model based on interval intuition fuzzy entropy and a dynamic attribute determination model based on game theory are proposed.Third,the inverse Poisson distribution method is used to assign weights to the time series.The information of target value assessment at different moments is effectively fused by constructing a dynamic fusion model of evidence inference methods in an interval intuitionistic fuzzy environment.Then the target value assessment results are obtained.The experimental simulation results show that the evaluation results of the proposed model have a smaller fuzzy interval length,thus reducing the uncertainty of the decision results compared with the traditional methods;meanwhile,the proposed model can meet the requirements of environmental dynamics.The proposed model can provide a reasonable and practical target value assessment for command information systems.(2)Research on target threat assessment with decision-oriented information in the form of possibility distribution: To address the problem of inaccurate target threat assessment results caused by the dynamic change trend of target detection information in a certain time period and the influence of the commander’s limited rationality on the target threat assessment results.A target threat assessment model based on the possibility distribution extended CPT-TODIM algorithm is proposed.First,the set of aerial target threat indicators is determined,considering various influencing factors affecting the target threat assessment.Using the possibility distribution to characterize the decision information in the target threat assessment and determining the initial decision matrix at different moments.Second,the inverse Poisson distribution method is used to assign time series weights to the decision information at different moments,and the initial decision matrices at different moments are assembled to obtain the integrated decision matrix.After that,a model for determining the weights of target threat assessment indicators based on the CRITIC algorithm is proposed,considering the influence of the degree of internal variability among target threat assessment indicators and the conflicting nature among indicators on the indicator weights.To measure the impact of the overall trend of target information on the target threat assessment,a possibility distribution mean matrix and a possibility standard deviation matrix were constructed.Target joint dominance can reflect the impact of a commander’s limited rationality on target threat assessment.Finally,by comparing and analyzing the simulation results derived from different threat assessment methods,the simulation results show that the model based on the possibility distribution extended CPTTODIM can fully reflect the change trend and fluctuation degree of target decision information in the time dimension and the influence of the commander’s limited rationality on the target threat assessment,which can realize the dynamic requirements of target threat assessment and can improve the reliability of the threat assessment results.(3)Research on combat plan evaluation with decision-oriented information in the form of trapezoidal neutrosophic number: In order to solve the problem that the decision weight information is completely unknown and it is difficult to reach a consensus decision due to the cognitive differences among commanders,a combat plan evaluation model based on the trapezoidal neutrosophic number extended MABAC algorithm is proposed.First,considering the uncertainty,indeterminacy,and inconsistency of information in a battlefield environment,the initial decision matrix is characterized by a trapezoidal neutrosophic number.Second,consistent assessment of information from different experts in decision-making on combat plans through consensus measurement and feedback mechanisms.In order to solve the problem of utterly unknown weight decision information in the evaluation of combat scenarios,the algorithm of TrNN-CRITIC is proposed to determine the attribute weights and the algorithm based on TrNN similarity measure to calculate the expert weights.Finally,Finally,by constructing the TrNN weighting matrix and the TrNN boundary approximation area matrix(BAA),a combat plan evaluation model based on the trapezoidal neutrosophic number extended MABAC algorithm is proposed.The simulation experiments and comparative analysis results show that the proposed algorithm can provide practical consensus evaluation of consistent combat scenarios when both attribute weights and expert weights are unknown.The research results of this paper enrich the theory and methods of assisted decisionmaking in uncertain environments,and the proposed model can provide a reference for key technologies of assisted decision-making in intelligent command and control systems.