Abstract
This article shows how restrictions of physical I/O
interfaces in terms of grouping, dual-power, and connector
segregation are considered correctly in the combinatorial
optimization of avionics architectures like large Integrated
Modular Avionics (IMA) systems. For each of the I/O sharing
restrictions a constraint formulation for multi-objective Integer
Linear Programing (ILP) is developed. The novel constraints are
applied for task assignment optimization in a verification example
and a real-world example of 536 tasks. Verification shows that the
I/O sharing is handled correctly and the real-world application
proofs the method to be industrially applicable. Moreover, the
influence on the solving speed is investigated. I/O restrictions can
rise the optimization time by factor 10, but the loss of the
optimization potential is below 10% for the tested objectives and
examples.
Users
Please
log in to take part in the discussion (add own reviews or comments).