Matchmaking Support for Mathematical Web Services

Matchmaking is one crucial tasks in agent-based systems and describes locating and identifying the most suitable services among all available services. This paper introduces matchmaking of mathematical services, whereby the matching is based on semantics using OpenMath object descriptions of preand post-conditions. A matchmaking architecture for mathematical services is described containing four matching algorithms which achieve a structural match, a syntax and ontological match, an algebraic equivalence match and a value substitution match. The matchmaking architecture calculates the match scores which give an indication of the quality of the matches. A case study explains in detail how the matching process for all four matching algorithms works.