RACSA - Rule Applying Comparison Shopping Agent


Fortgeschrittenenpraktikum SS 2002


Universität des Saarlandes Deutsches Forschungszentrum für künstliche Intelligenz
by Sabrina Scherer and Benjamin Olschok

(Sabrina dot Scherer at dfki dot de) (Benjamin dot Olschok at dfki dot de)

in cooperation with Dr. habil. Harold Boley (DFKI)




The motivation for RACSA comes from the lecture "Semantisches Web und Wissensmanagement" given by Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster at Universität des Saarlandes. Therefore our project contributes to the Semantic Web, especially to e-commerce in the Semantic Web.

Unlike in standard comparison shopping, the main idea of RACSA is marking up shopping web pages with Semantic Web languages and, based on their list prices, computing the real end prices by applying various business rules.

From a more technical point of view, RACSA is a rule empowered shopping agent, using RuleML as the XML-based rule language and Mandarax (a Java-based Prolog engine by Jens Dietrich) as the inference component for processing knowledge and data bases (i.e. facts and rules).

The goal of RACSA is to demonstrate an e-commerce application enabled by the Semantic Web techniques: Supporting purchase-decisions by simplifying price comparisons.

We restrict our project to an exemplary domain: Refrigerators/freezers offered by 3 small fictitious shops. Rule application takes place in a wide variety of areas, such as gross/net prices, calculation of warranty time, additional insurance and transport costs, payment and customer discounts.

As we develop RACSA within the scope of a project at Universität des Saarlandes (a so-called "Fortgeschrittenenpraktikum") and DFKI, we do not pursue any commercial goals with RACSA online shopping.




see instructions for use first