基于半模型的OWL本体理由探求方法研究  

Semi-Models Based Justifications Detection for OWL Ontologies

作　　者：张瑜[1,2] 欧阳丹彤[1,2] 崔仙姬[3] 叶育鑫 ZHANG Yu;OUYANG Dan-Tong;CUI Xian-Ji;YE Yu-Xin(College of Computer Science and Technology,Jilin University,Changchun 130012;Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education,Changchun 130012;College of Information and Communication Engineering,Dalian Minzu University,Dalian,Liaoning 116600)

机构地区：[1]吉林大学计算机科学与技术学院,长春130012 [2]符号计算与知识工程教育部重点实验室,长春130012 [3]大连民族大学信息与通信工程学院,辽宁大连116600

出　　处：《计算机学报》2018年第12期2720-2733,共14页Chinese Journal of Computers

基　　金：国家自然科学基金项目(61672261;61502199)资助~~

摘　　要：OWL本体理由探求是语义Web推理的重要任务之一.随着语义Web数据的急剧增长以及本体规模的不断扩大,目前的本体理由探求策略已难以满足它们对推理性能的要求.该文以基于黑盒的探求技术为研究对象,黑盒法是基于"扩张"和"收缩"两个阶段实现理由探求任务的,"扩张"阶段的目标是获得蕴涵目标公理的理由的一个超集,"收缩"阶段对得到的理由超集进行删减至极小集合.然而,这两个阶段的主要时间开销在于频繁地调用推理机进行变化的公理集合与目标公理之间的蕴涵关系的检测,这会严重影响理由探求的效率.为了解决这一问题,通过观察理由探求过程中公理集合的变化情况,给出增量本体序列定义,并揭示了增量本体序列中的最大增量本体与理由之间的关系.增量本体序列的生成过程主要涉及两方面因素:(1)后继本体对先驱本体的有效扩充(必须保证是拟序关系);(2)对本体链中的本体是否蕴涵目标公理的推理判定(必须保证当且仅当最大本体蕴涵目标公理).在增量本体序列生成过程中,利用半模型证明了后继增量本体与目标公理之间的蕴涵关系是半可判定的,进而给出基于半模型增量推理的理由超集探求算法及其正确性证明.半模型增量推理的增量体现在:保留上一次得到的模型作为下一次判定的初始条件之一,从而避免传统蕴涵判定中,每一次都完全重构模型的冗余计算.最后,提出了一种与现有的收缩过程相反的理由求解方案——基于扩张的理由求解策略.通过迭代地添加公理过程,探测该公理集下的所有理由的公共元素.利用探测到的所有公共元素构造目标公理的理由.随后利用该文提出的增量推理任务分别给出新的"扩-缩"理由探求方法和"双扩"理由探求方法.实验结果表明,改进后的"扩-缩"理由探求方法在求解性能上优于原有的"扩-缩"理由探求方法;而新提�Detecting justifications for OWL ontologies is an important task in semantic web reasoning. Because of the rapid growth of semantic web data and the continuous increase of scales of ontologies, existing semantic web reasoning strategy has been difficult to adapt to such environment. In this paper, we discussed black-box techniques of detecting justifications. Black box method for detecting justifications is based on the “expansion” stage and “contraction” stage. The “expansion” stage can achieve a superset of justification which entails the target axiom. The “contraction” stage prunes the superset to obtain a minimal set of axioms. However, the main cost of it is the frequent call to the ontology reasoners to determine the implication relationship between changing axioms and target axioms, thus affecting the efficiency of the detecting. In order to solve this problem, the definition of the incremental ontologies sequence is given and the relationship between the maximum of incremental ontology and the justification is revealed base on the observation of the changes of axiom sets. The generation of incremental ontologies sequence is mainly involves two important factors： （1 the effective expansion from the successor ontology to the precursor ontology （it is necessary to ensure that this expansion must be a quasi-order relation, （2 the reasoning judgment on whether the ontology in the ontology chain entails the target axiom. During the generation of the incremental ontologies sequence, a new reasoning task, incremental reasoning task, is discovered. Furthermore, semi-decision between next incremental ontology and target axiom is proven by semi-model which definition is based on the current incremental ontology and its model. It leads to the algorithm about detecting the super-set of justification, and the correction of this algorithm is proven. Semi-model incremental reasoning embodies in： reserves the model obtained in the last time as one of the initial conditions, so as to avoid the com

关 键 词：语义Web推理 OWL本体 理由 黑盒技术 蕴涵关系 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]