Research On Logical Contradiction Elimination Method Based On Cooperative Game And Its Application

The ontology knowledge base based on description logic is an important method for the Semantic Web to achieve the domain knowledge sharing.However,in the process of constructing,merging and updating ontology knowledge bases,the unavoidable logical conflicts lead to the loss of reasoning ability of ontology knowledge bases,and prevents domain knowledge sharing.The classical logical conflict elimination method is based on the Reiter’s hitting set tree algorithm,which calculates minimal hitting sets that hit with all minimal conflict sets,and removes one of minimal hitting sets in the formula set to eliminate logical conflicts.In order to improve the efficiency of conflict elimination based on the hitting set tree method,researchers use mathematical programming thinking,construct a linear programming model to describe conflict elimination,and calculate the minimum hitting sets（the set with the smallest number of elements in minimal hitting sets）while following the minimal change principle.However,the linear programming based approach to eliminate logical conflicts treats the formulas equally,which results in a large number of model solution sets.In this dissertation,we introduce the Shapley Value to inscribe the formula inconsistency degree,mine the logical conflict nature of the ontology knowledge base,and use the multi-objective optimization technology to reduce the number of solution sets effectively,the main research content and innovation points are as follows.（1）According to the study of inconsistency measure based on the Shapley Value to measure the contribution of deleted formulas to logical conflict elimination,we calculated the Shapley marginal contribution value reflecting the degree of formula inconsistency,as well as the priority of deleting formulas in logical conflict elimination.（2）We constructed a multi-objective linear programming model that eliminates logical conflicts using multi-objective planning approach.The number of solution set bases is first restricted using the base of the minimum hitting sets,and then modeled using the priority reflected by Shapley marginal contribution value to rationally select the higher inconsistency formula set as the optimal solution set,while following the minimal change principle,and reducing the solution set number of the single-objective solution pool.（3）We explored the similarity between logical conflict elimination and smart distribution grid for smart power system.Moreover,according to the actual distribution grid application scenario,we constructed the power programming model for distribution grid based on the logic conflict elimination thought,and verified the model complexity by actual engineering application.