A Weighted-Tree Similarity Algorithm for Multi-Agent Systems in E-Business Environments

A tree similarity algorithm for match-making of agents in e-business environments is presented. Product/service descriptions of seller and buyer agents are represented as nodelabelled, arc-labelled, arc-weighted trees. A similarity algorithm for such trees is developed as the basis for semantic match-making in a virtual marketplace. The trees are exchanged using a serialization in Object-Oriented RuleML. A corresponding Relfun representation is employed to implement the similarity algorithm as a parameterised, recursive functional-logic program. Results from our experiments are found to be meaningful for e-business/e-learning environments. The algorithm can also be applied in other environments wherein weighted trees are used. Keywords/phrases: multi-agent system, e-business, e-learning, similarity measure, arclabelled trees, arc-weighted trees, Object-Oriented RuleML, Relfun.