Membrane Separation Processes Using Machine Learning Based Mathematical Programming Models - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Membrane Separation Processes Using Machine Learning Based Mathematical Programming Models

Membrane separation processes are a rapidly spreading technology, since they can result in energy efficient, small sized and environmental friendly processes. Their optimal design can be modelled as a MINLP problem. The resulting problem can be reduced to a small subset of NLP subproblems, under some restrictive hypotheses, and using symmetry reduction techniques. Nevertheless, each of these sub-problems remains challenging due to the large number of non-convex equality constraints that represent the physical behaviour of each membrane in the system. The aim of this phase of our work is to compare the original full-equation model with a machine learning hybridated model (used to represent the single membrane behaviour) in terms of quality of solutions and computational times

Amalia Macali
  • Fonction : Auteur
Christophe Castel
Veronica Piccialli
  • Fonction : Auteur
Alice Calamita
  • Fonction : Auteur
Francesco Di Luzio
  • Fonction : Auteur
Fichier non déposé

Dates et versions

hal-03902346 , version 1 (15-12-2022)

Identifiants

  • HAL Id : hal-03902346 , version 1

Citer

Amalia Macali, Bernardetta Addis, Christophe Castel, Veronica Piccialli, Alice Calamita, et al.. Membrane Separation Processes Using Machine Learning Based Mathematical Programming Models. informs, Jan 2020, online, France. ⟨hal-03902346⟩
25 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More