基于部分参考匹配集合的混合遗传算法解决本体匹配问题

Optimizing ontology matching through hybrid genetic algorithmbased on partial reference alignment

不同本体之间的异构问题成为各种应用之间实现更智能化、高效的知识共享和通信的障碍。本体匹配是解决上述问题的有效方法。为了获取高质量匹配结果,提出了基于部分参考匹配结果(partial reference alignment, PRA)的混合遗传算法,该方法采用分层选择方法解决传统PRA构建过程中的语义丢失问题,并提出了一种新的适应度函数进一步充分利用PRA中的潜在信息,从另一个角度解决语义丢失问题。此外,该算法结合了遗传算法(GA)和随机爬山算法两种方法,以在全局和局部范围内寻找最优的本体匹配方案。实验结果表明,该算法在不同的本体匹配任务中均能有效地获得高质量的匹配结果,并且和其他前沿的方法比较也有出色的表现。

The problem of heterogeneity between different ontologies becomes an obstacle to more intelligent and efficient knowledge sharing and communication between various applications.Ontology matching is an effective way to solve the above problems.In order to obtain high quality matching results,this paper proposed a hybrid genetic algorithm(HGA)based on PRA.The method adopted a stratified selection approach to utilize the heterogeneity feature among ontologies to solve the issue of semantic loss in the traditional PRA construction process,and proposed a new fitness function to further fully utilize the potential information in the PRA to solve the semantic loss problem from another perspective.In addition,the algorithm combined both genetic algorithm and stochastic hill climbing algorithm in order to find the optimal ontology matching solution in both global and local scales.Experimental results show that the algorithm is effective in obtaining high-quality matching results in different ontology matching tasks,and it also performs well in comparison with other cutting-edge methods.