Matching Sensor Ontologies Through Compact Evolutionary Tabu Search Algorithm

Abstract

Although sensor ontologies are regarded as the solution to data heterogeneity on the Semantic Sensor Web (SSW), these sensor ontologies themselves introduce heterogeneity by defining the same entity with different names or in different ways. To solve this problem, it is necessary to determine the semantic identical entities between heterogeneous sensor ontologies, so-called sensor ontology matching. Due to the complexity of the sensor ontology matching process, Evolutionary Algorithm (EA) can present a good methodology for determining ontology alignments. To overcome the EA-based ontology matcher’s shortcomings, i.e. premature convergence, long runtime and huge memory consumption, this paper present a Compact Evolutionary Tabu Search algorithm (CETS) to efficiently match the sensor ontologies. The experiment utilizes Ontology Alignment Evaluation Initiative (OAEI)’s bibliographic benchmark and library track, and two pairs of real sensor ontologies test CETS’s performance. The experimental results show that CETS is both effective and efficient when matching ontologies with various scales and under different heterogeneous situations, and comparing with the state-of-the-art sensor ontology matching systems, CETS can significantly improve the ontology alignment’s quality.