The automatic composition and execution of workflo ws composed of web and grid services is nowadays a challenging tas k that is hardly achievable. One of the popular solutions is to utilize semantic annotations to the services and apply them during the composition phase. Our ap proach is based on utilization of not only the semantic annotations of services, but also the results produced in previous executions of workflows and th eir meta-information in form of QoS (Quality of Service) parameters. This i nformation helps us to minimize the human participation that is left only for the reasons of having the chance to involve directly domain experts and to be tter control the execution phase of workflows. The workflow is modeled as a Pe tri net, which allows for easy incorporation of data into the process and uti lizes semantic description of resources by means of ontologies 1.
[1]
Ladislav Hluchý,et al.
Knowledge-Based Runtime Prediction of Stateful Web Services for Optimal Workflow Construction
,
2005,
PPAM.
[2]
M. Sarnovsky,et al.
Text mining workflows construction with support of ontologies
,
2008,
2008 6th International Symposium on Applied Machine Intelligence and Informatics.
[3]
Viera Rozinajová,et al.
Bridging the Gap Between Service-Oriented and Object-Oriented Architectures in Information Systems Development
,
2008,
Advances in Information Systems Research, Education and Practice.
[4]
Ulf Leser,et al.
Querying Distributed RDF Data Sources with SPARQL
,
2008,
ESWC.
[5]
Katarzyna Rycerz,et al.
Grid Service Registry for Workflow Composition Framework
,
2004,
International Conference on Computational Science.