Research On The Framework And Algorithm Of Uncertainty Elimination Based On Multiple Criteria Decision Making

In the research and application process of context-aware systems(CASs),it is inevitable to face the problem of uncertainty of context information.However,the traditional methods of dealing with uncertainty problems generally have the disadvantages of poor flexibility,high redundancy,and low degree of collaboration,which cannot accurately and reasonably deal with useful information,and cannot be treated according to the symptoms.Therefore,how to achieve different types and scenarios context information classification processing to meet users' requirements for system universality and flexibility is an important aspect of the uncertainty of context information.In addition,for the inconsistency problem that is the most difficult to resolve in the uncertainty processing,the traditional inconsistency processing algorithm has the problems of low flexibility and low accuracy,so improving the shortcomings of traditional inconsistency processing algorithms is also a top priority in the research of context information.Aiming at the above problems,this thesis mainly conducts in-depth research from the following aspects:(1)Aiming at the problems of traditional context-aware systems in dealing with uncertainty,this thesis proposes a multi-source context uncertainty processing system and the corresponding algorithm based on multiple criteria decision making.The algorithm selects the most effective processing strategy in real time based on the state of system operation and the type of context,and for the different types of context,the advantages of multiple processing strategies are integrated.The algorithm can well meet the universality requirements in context information processing,and can be flexibly and conveniently applied to the uncertainty information processing of context information of different types and different environments.(2)Aiming at the problems of the traditional context information inconsistency elimination algorithm,this thesis proposes an inconsistency elimination algorithm based on set pair analysis(SPA)and QoX quality parameters.The algorithm uses some parameters of the QoX to quantify the quality of the context information collected by the context sources,which improves the completeness and accuracy of the quantification of the context information quality.In addition,the algorithm uses SPA to perform a hierarchical analysis of the required quality parameters,and users subjective weights are used to assign weights to the degree of connection after the analysis of set pairs,taking into account the flexibility and accuracy in the process of calculating the quality of context information.With the background of wearable health monitoring system,simulation experiments are conducted from various aspects to verify the effectiveness of the proposed algorithms.Through the above research,on the one hand,the processing methods that can effectively solve the shortcomings of traditional uncertainty processing methods such as poor flexibility,high degree of redundancy and low degree of cooperation,and realize reasonable and effective uncertainty and inconsistency processing of useful context information.On the other hand,it is possible to comprehensively consider the various context quality parameters of the data source and the subjective requirements of the users to achieve high-precision processing of the context information.It has laid a good foundation for the subsequent application of context-aware information,improved the overall performance of context-aware systems,and provided strong support for the development and wide application of context-aware technology.