Brokering Web Services via a Hybrid Ontology Mediation Approach Using Multi Agent Systems (MAS)

Brokering Web Services via a Hybrid Ontology Mediation Approach Using Multi Agent Systems (MAS)

Saravanan Muthaiyah (Multimedia University, Malaysia) and Larry Kerschberg (George Mason University, USA)
DOI: 10.4018/978-1-60566-910-6.ch023
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

This chapter introduces a hybrid ontology mediation approach for deploying Semantic Web Services (SWS) using Multi-agent systems (MAS). The methodology that the authors have applied combines both syntactic and semantic matching techniques for mapping ontological schemas so as to 1) eliminate heterogeneity; 2)provide higher precision and relevance in matched results; 3) produce better reliability and 4) achieve schema homogeneity. The authors introduce a hybrid matching algorithm i.e. SRS (Semantic Relatedness Score) which is a composite matcher that comprises thirteen well established semantic and syntactic algorithms which have been widely used in linguistic analysis. This chapter provides empirical evidence via several hypothesis tests for validating our approach. A detailed mapping algorithm as well as a Multi-agent based system (MAS) prototype has been developed for brokering Web services as proof-of-concept and to further validate the presented approach. Agent systems today provide brokering services that heavily rely on matching algorithms that at present focus mainly only on syntactic matching techniques. The authors provide empirical evidence that their hybrid approach is a better solution to this problem.
Chapter Preview
Top

The Problem

This section highlights the interoperability problem. The main reasons for data heterogeneity amongst ontologies are: 1) structural heterogeneity (difference taxonomy structures); 2) semantic data heterogeneity (scale and representation conflict); 3) subjective mapping (conflicting data instances) and 4) atomic stored data (conflicting data type value) (Stuckenschmidt, Wache, & Visser). Sources for semantic heterogeneity also include differences in data-definition constructs, differences in object representations, and system-level differences in the way that atomic data (e.g., byte order for multibyte data, such as an integer) is stored in the two systems (Maedche, Motik, & Stojanovic, 2002).

Complete Chapter List

Search this Book:
Reset