Adaptation by Nash games in gradient-based multi-objective/multi-disciplinary optimization - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

Adaptation by Nash games in gradient-based multi-objective/multi-disciplinary optimization

Abstract

A two-phase numerical process is proposed for gradient-based multidisciplinary optimization. In a first phase, one or several Pareto-optimal solutions associated with a subset of the cost functions, the primary objective functions, subject to constraints, are determined by some effective multi-objective optimizer. In the second phase, a continuum of Nash equilibria is constructed tangent to the Primary Pareto Front along which secondary cost functions are to be reduced, while best preserving the Pareto-optimality of the primary cost functions. The focus of the article is on estimating the rate at which the secondary cost functions are diminished. The method is illustrated by the numerical treatment of the optimal sizing problem of a sandwich panel w.r.t. structural resistance under bending loads and blast mitigation.
Fichier principal
Vignette du fichier
desideri-jano13.pdf (691.19 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03430972 , version 1 (16-11-2021)

Identifiers

Cite

Jean-Antoine Désidéri. Adaptation by Nash games in gradient-based multi-objective/multi-disciplinary optimization. JANO13 - Mathematical Control and Numerical Applications, Feb 2021, Khouribga, Morocco. ⟨10.1007/978-3-030-83442-5_9⟩. ⟨hal-03430972⟩
65 View
77 Download

Altmetric

Share

Gmail Facebook X LinkedIn More