基于本体的模型集成  

Ontology Based Model Integration

作　　者：蔡树彬[1,2] 明仲[2] 李师贤[1] 刘显明 

机构地区：[1]中山大学计算机科学系,广东广州510275 [2]深圳大学计算机与软件学院,广东深圳518060 [3]江西电力信息通讯有限公司,江西南昌330077

出　　处：《电子学报》2009年第4期713-719,共7页Acta Electronica Sinica

基　　金：国家自然科学基金(No.60673122);江西省电力公司科技项目(No.200850802);广东省自然科学基金(No.7301315,No.815103000700002)

摘　　要：模型在决策支持系统中具有非常重要的作用.如何更好地集成模型,实现模型协作,一直是该领域的重点研究内容.模型参数的表示和匹配是模型集成的核心问题.由于难以描述模型输入、输出参数的语义,模型集成的自动化程度、智能化程度、扩展性、开放性和可理解性等性能一直不够理想.提出一种基于本体的程序-逻辑混合(PLM)模型表示法.该方法易于实现,支持推理,具有高开放性和可扩展性.使用本体描述模型参数的语义,并基于参数的语义匹配,设计模型集成策略,可自动、智能地创建易于理解的模型集成方案.给出模型集成的形式化定义并证明相关定理.Model plays very important role in Decision Support System（DSS）. How to integrate models and improve models collaboration have always been the key research spot in DSS field. Model parameters＇ representation and matching are the core problems in model integration Due to the difficulties in describing the semantic of the input,output parameters,the intelligence,extonsibility, openness and understanding of automatic model integration are not satisfied and need to be improved. An ontology-based program-logic-mix（PLM）model representation is proposed, which is easily-implemented, reasoning-supported, has high openness and extonsibility. Ontology is used to describe the semantics of parameters. Based on this semantic match, model integration strategies are designed to create understandable decision problem solutions. Formal definition of model integration is given and relevant theorems are proved.

关 键 词：决策支持系统 模型集成 模型表示 本体 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]