1424. Logic-Based Benders Decomposition for Joint Routing and Scheduling of TSN networks in Avionics
Invited abstract in session MC-7: Mathematical Programming in Scheduling, stream Scheduling and Project Management.
Monday, 12:30-14:00Room: Clarendon GR.01
Authors (first author is the speaker)
| 1. | Jesper Vines
|
| Department of Mathematics, Linköping University, Saab AB | |
| 2. | Hannes Uppman
|
| Saab AB | |
| 3. | Mikaela Lindberg
|
| Saab AB | |
| 4. | Elina Rönnberg
|
| Department of Mathematics / Optimization, Linköping University |
Abstract
In the electrical system in an aircraft, the avionic system, a large number of data packages need to be reliably transmitted and received on time. This can be achieved by jointly deciding how to route packages through the network and determining a schedule for the send times of packages. The recently proposed standards of Time-Sensitive Networking enable guarantees regarding latency and availability of such scheduled data. However, a challenging optimisation problem for the routing and scheduling needs to be solved. The expected increase in avionic system sizes, and the number of data packages sent, require the development of optimisation methods that scale well.
We propose a Logic-Based Benders Decomposition scheme for joint routing and scheduling in a TSN network. There is a natural decomposition of this problem, where message routing is decided in the master problem and the remaining scheduling problem is solved in the subproblem. With this decomposition, the scheduling subproblem can become challenging to solve for larger problem instances. Therefore, we strengthen the master problem by taking scheduling partially into account at the outset, for example by considering network load or transmission delays. Through this, the size of the subproblem is reduced and larger problem instances can be solved.
Keywords
- Engineering Optimization
- Programming, Constraint
- Programming, Mixed-Integer
Status: accepted
Back to the list of papers