OTAGen
On the one hand, agent-based software platforms are commonly used these days, while on the other hand Semantic Web technologies are also maturing. It is obvious that the combination of these two technologies can bring added value and create a Semantic Agent-based framework. However, it also known that these Semantic Web technologies, and the reasoning on ontologies in particular, can rapidly become resource intensive. In order to get a clear view on this problem, we have developed OTAGen, a highly tunable tool to generate ontologies and corresponding queries. The generated ontologies can then be used to benchmark Semantic Applications, and to define a suitable size and complexity so that the agents can still handle the model.
Few ontology generators have been developed previously, with the aim to benchmark Semantic Applications and profile their behaviour with different sizes and complexity of the used ontology. Arguably the most wellknown is the Lehigh University Benchmark (LUBM) (http://swat.cse.lehigh.edu/projects/lubm/). The LUBM features a university domain ontology, of which the T-Box is statically defined. Also a set of 14 different queries on this data is included in the benchmark. The size of the dataset, the A-Box, can be specified and varied during the generation process of the actual ontologies. The perfomance of Semantic Applications can be measured according to the behaviour of the application, processing the predefined queries on the different generated ontologies.
An extension of LUBM, namely the University Ontology Benchmark (UOB) has been presented in http://www.springerlink.com/content/l0wu543x26350462/. After all, one of the big disadvantages of LUBM is the statically defined T-Box. UOB aims to overcome this problem, by introducing OWL-Lite and OWL-DL constructs in the LUBM T-Box, thus creating extra parameters to be varied. The performance of Semantic Applications can again be measured according to the characteristics of the generated ontologies. However, still a more or less static T-Box is used, although completed with additional OWL-Lite and OWL-DL constructs. Our work with OTAGen aims to provide the capibility of specifying a large range of parameters characterising the ontology, both on T-Box as well as A-Box level. OTAGen also generates queries, of which the characteristics can be specified.
Publications>
Download
You can download below an alpha-version of the OTAGen tool. This tool is released under AGPL license. You also need the third party Jena library for this tool to work. Jena can be downloaded seperately from: http://jena.sf.net/
You need a recent version of Java (preferrably 1.5 or higher) and the tool can be started with the command: java -jar OTAGen.jar. For more information on the available configuration parameters, we refer to one of the publications listed above. As mentioned, this tool is alpha quality. Therefore, bugs can be expected and we appreciate all feedback. Feature requests and bug-reports can be mailed to: otagen@intec.UGent.be.