Design und Management von Experimentier-Workflows

Abstract

Experimentation in my work means performing experiments based on computer-based models, which describe system structure and behaviour abstractly. Often, these systems do not exist in reality yet or can not be explored directly for other reasons. Hence, instead of the system itself models of the system will be explored (with a certain goal). These models are executed in special computing environments (usual simulation systems) to get model observations. Systematic experimentation using model input parameter variation assignments leads to lots of possibly long-running experiments that must be planned, documented, automated executed, monitored and evaluated. The problem is, that experimenters (who are usually not computer scientists) miss the proper means of expressions (e. g., to express variations of parameter assignments) to describe experimentation processes formally in a way, that allows their automatic execution by a computer systemwhile preserving reproducibility, re-usability and comprehension (summarized as transparency). An existing approach to tackle this problem is using a scientific workflow management system (S-WfMS), which is applicable, since the process of experimentation can be defined as a specifically structured scientific process that comprises the ordered phases planning, execution and evaluation of experiments. A formalization of such a process is further called as experimentation workflow (a specialization of a scientific workflow). A S-WfMS is flexible enough to handle experimentation workflows, while using only the concepts of data and control flow. It contains no knowledge about the semantics of the data and how it is related to the model or the experimentation aim. This is problematic, because experimenters are focused on the aim. My approach is to identify general experimentationworkflow concepts and formalize them as a meta-model-based domain-specific language (DSL) that I call experimentation language (ExpL). experimentation language (ExpL) includes general workflow concepts like control flow and the composition of activities, and it allows modeling of experimentation workflows on a framework-independent, conceptional level. This approach is based on the idea to apply the paradigm of model driven engineering also to the domain of experimentation workflows. Hence, re-using and sharing the experimentation workflow with other scientists is not limited to a particular framework anymore. Therefore, it is easier to switch the particular framework that executes the model because the experimentation description does not need to be re-implemented. This is useful in order to follow different experimentation aims if the type of the model supports switching the framework. ExpL includes common workflow concepts like activities and resources, but provides declarative elements in order to support parameter studies properly, too. This is a novelty that helps

Extracted Key Phrases

Cite this paper

@inproceedings{Khnlenz2014DesignUM, title={Design und Management von Experimentier-Workflows}, author={Frank K{\"{u}hnlenz}, year={2014} }