Master's Thesis   
 

Translating GPSS to DEVS (Based on a DEVS Building Block Library)

Abstract

Models are everywhere. From simple mathematical equations to a schematic representation of a water filtering plant to a waiting line at the cash registry in a super market. They're often used to get a solid understanding of business flows, system flows, proof-of-concepts, factories... Different modeling languages within the domain of Modeling and Simulation (M&S) can be used for various purposes. A subset of these languages describes discrete-event systems and can be further separated based on their world view: event scheduling, activity scheduling or process interaction. This thesis will use the modeling formalism of Discrete EVent Systems (DEVS), a general-purpose event-scheduling language, to create a generic building block library (BBL). Multiple tools, that are currently used for big systems by major companies, will be explored as a foundation for this library. Additionally, the General Purpose Simulation System (GPSS), an example language in the process interaction world view, will be described and a subset thereof will be translated onto the DEVS formalism. This way, we benefit from the advantages that DEVS offers, while still enabling the strengths of GPSS. Accompanied with this paper, there is a Python-implementation (based on the Python(P)DEVS kernel) of the building blocks and the translation that will be discussed in these pages.

Links

Thesis Report
The thesis report can be downloaded here.
Powerpoint from Intermediary Conference
The google presentation is located here.
Presentation from Evaluation
The google presentation can be found here.
Mindmap
Besides being available below, you can also access it on MindMeister.
Code and Documentation
The code can be found on the MSDL git server. The documentation is located in docs/_build/html.
Notes on reading
Additional notes and bundlings of certain data can be found on this page. Yet, in general, it is better to read the thesis instead.
Maintained by Randy Paredis.