Research On The Semantic Driven Conflict Management Of Enterprise Collaborative Modeling

Domain experts master different aspects of information and requirements about the enterprise operations. In order to ensure that the enterprise models could comprehensively and precisely reflect the actual situation of the enterprise operations and business requirements in the future, domain experts have to collaborate to build enterprise models. Under support from the Ministry of Railway-Tsinghua science and technology research program (No. t200412) and the National High Technology Research and Development Program (No. 2008AA04Z102), this dissertation conducts a systematic research on the methods and relative theories of semantic driven conflict management of enterprise collaborative modeling. The methods presented by this dissertation could realize the transformation between models of different modeling methods, standardize the descriptions of model elements, support the match between elements from different models, automatically detect the conflicts between models built by different domain experts, and integrate models to get relatively complete models.In order to endow models with formal semantics, this dissertation proposes a decidable knowledge representation, Horn-Extended DL, which combines the representation and reasoning abilities of Description Logic and Logic Program, develops the Tableau algorithm to support the automatic reasoning of Horn-Extended DL, and proves the termination, soundness and completeness of the Tableau algorithm.This dissertation puts forward a model transformation framework based on Horn-Extended DL. With the representation and reasoning abilities of Horn-Extended DL, the syntax correctness of models could be verified, the explicit knowledge input by domain experts not only could be retrieved as the base for new models, but also could be reasoned to get implicit knowledge, so that the model knowledge's value and use efficiency are improved.Due to different language habits, domain experts may use different terms to describe model elements. In order of resolve the conflicts of model elements' descriptions, this dissertation proposes to standardize the elements'descriptions and ensure their definite semantic meaning by semantic feature decomposition and limited word set, develops similarity calculation and search mechanism for limited words and concepts to support the reuse of the model elements'descriptions and the match between elements from different models.In order to ensure the model's correctness and completeness, this dissertation develops two model operations: comparison and integration. By comparing the models built by different domain experts, conflicts between models could be automatically detected. Based on the conflicts found, domain experts could collaborate to reach agreement on the models. By integrating incomplete models, relatively complete models could be got.