|
Title
|
|
|
|
Hint-based configuration of co-simulations with algebraic loops
| |
|
Author
|
|
|
|
| |
|
Abstract
|
|
|
|
| Co-simulation is a powerful technique for performing full-system simulation. Multiple black-box models and their simulators are combined together to provide the behaviour for a full system. However, the black-box nature of co-simulation and potentially infinite configuration space means that configuration of co-simulations is a challenging problem for today’s practitioners. Our previous work on co-simulation configuration operated on the notion of hints, which allow system engineers to encode their knowledge and insights about the system. These hints, combined with state-of-the-art best practices, can then be used to semi-automatically configure the co-simulation. We summarize our previous hint-based configuration work here, and explore the challenging problem of scheduling co-simulations which contain algebraic loops. Solving or “breaking” these loops is required for scheduling, yet this breaking process can induce errors in the co-simulation. This work formalizes this scheduling problem, presents our insights gained about the problem, and details an optimal search algorithm as well as greedy scheduling algorithms. These heuristic algorithms are evaluated on (synthetic) co-simulation scenarios to determine their relative speedup and optimality. |
| |
|
Language
|
|
|
|
English
| |
|
Source (book)
|
|
|
|
Simulation and modeling methodologies, technologies and applications : 9th International Conference, SIMULTECH 2019, Prague, Czech Republic, July 29-31, 2019: revised selected papers
| |
|
Source (series)
|
|
|
|
Advances in intelligent systems and computing ; 1260
| |
|
Publication
|
|
|
|
Cham
:
Springer
,
2020
| |
|
ISBN
|
|
|
|
978-3-030-55866-6
| |
|
DOI
|
|
|
|
10.1007/978-3-030-55867-3_1
| |
|
Volume/pages
|
|
|
|
p. 1-28
| |
|
Full text (Publisher's DOI)
|
|
|
|
| |
|