|
Title
|
|
|
|
Property preservation in co-simulation
| |
|
Author
|
|
|
|
| |
|
Abstract
|
|
|
|
| Modeling and Simulation (M&S) techniques are today extensively used both in industry and science, to develop and understand complex systems. As development processes morph to respond to economic pressures, they impose new demands on these techniques: - frequent full system evaluation is required to prevent late integration problems among specialized teams who worked in parallel on different, but interconnected, parts of the system; and - seamless integration of externally supplied component models into the M&S workflow of Original Equipment Manufacturers is needed to enable high fidelity simulations. Traditional M&S techniques, where a single model of the whole system is built and simulated, are insufficient to address these demands, because: - teams use mature M&S tools, each tailored to a particular domain, and not capable of exporting models that are compatible with any other M&S tool; and - external suppliers are not willing to share high fidelity models without expensive contracts protecting their Intellectual Property (IP). Co-simulation is a way to tackle these challenges. It consists of the theory and techniques to enable global simulation of a coupled system via the composition of simulators. Each simulator is broadly defined as a emph{black box} capable of exhibiting behavior, consuming inputs and producing outputs. This nature is also what aggravates the fundamental challenge in co-simulation: deciding whether the results can be trusted. This thesis is comprised of two parts. The first part explores the challenge, and tries to understand what makes co-simulation different than traditional simulation techniques. One of the conclusions of this part is the need to ensure that co-simulations preserve properties of the system being developed (e.g., stability, smoothness, etcldots). The second part represents a collection of work, each targeting an aspect of property preservation, including giving the users of co-simulation, the ability to control the implementation of participating simulators, and then providing a framework to help them do so correctly. The reported results provide a deeper understanding of the fundamental challenges that need to be addressed before co-simulation can become a seamless technique in the development of complex systems, including: - how to configure adaptive co-simulation algorithms that preserve stability; - how to configure state event location for co-simulation of hybrid systems; - a tool that allows the configuration of simulators participating in a co-simulation; and - a framework that guides the configuration of the co-simulation, so as to preserve domain specific description of system properties. |
| |
|
Language
|
|
|
|
English
| |
|
Publication
|
|
|
|
Antwerpen
:
Universiteit Antwerpen, Faculteit Wetenschappen, Departement Informatica
,
2019
| |
|
Volume/pages
|
|
|
|
229 p.
| |
|
Note
|
|
|
|
:
Vangheluwe, Hans [Supervisor]
:
De Meulenaere, Paul [Supervisor]
| |
|
Full text (open access)
|
|
|
|
| |
|