Evidence-based robust optimisation of space systems with evidence network models

Filippi, Gianluca and Marchi, Mariapia and Vasile, Massimiliano and Vercesi, Paolo; (2018) Evidence-based robust optimisation of space systems with evidence network models. In: 2018 IEEE Congress on Evolutionary Computation. IEEE, BRA. ISBN 9781509060177 (https://doi.org/10.1109/CEC.2018.8477917)

[thumbnail of Filippi-etal-IEEE-2018-Evidence-based-robust-optimisation-of-space-systems-with-evidence]
Preview
 Text. Filename: Filippi_etal_IEEE_2018_Evidence_based_robust_optimisation_of_space_systems_with_evidence.pdf
Accepted Author Manuscript
Download (468kB)| Preview

Abstract

The paper presents an approach to optimise complex systems in space systems engineering, accounting for epistemic uncertainty. Uncertainty is modelled with Dempster-Shafer theory of Evidence and the space system as a network of connected components. A constrained min-max problem is then solved, with a memetic algorithm, to deliver a robust design point. Starting from this robust design point a sequence of evolutionary optimisation steps are used to reconstruct an approximation of the Belief and Plausibility curves associated to a particular design solution. The constrained min-max approach and the evolutionary reconstruction of the Belief and Plausibility curves are tested on one realistic case study of space systems engineering.

CORE (COnnecting REpositories)