Skip to main content
Vai all'homepage della Commissione europea (si apre in una nuova finestra)
italiano it
CORDIS - Risultati della ricerca dell’UE
CORDIS
Contenuto archiviato il 2024-05-07

High-performance interior-point based decomposition methods for structured convex optimization

Obiettivo



Research objectives and content
The purpose of the present project is to study interior-point based decomposition methods for structured convex optimization. Indeed in this project we will develop an interior-point based decomposition method which has good | theoretical properties and makes it possible to solve larger structured convex optimization problems than using | the current available techniques. Hence, in this project we also implement and test the proposed method on a computer. Finally, we will also investigate parallelization of the proposed method. This may lead to a very | powerful method for solution of certain classes of problems such as multi-commodity network flows.
Training content (objective, benefit and expected impact)
The traning component in the project is that the applicant learns about decomposition methods and notably interior-point based decomposition methods and their efficient implementation. This will allow the applicant to do | research on decomposition methods in his future career as a researcher.
Links with industry / industrial relevance (22)
With respect to industry involvement in the project, then the applicant has contact to Dash Associates U.K. which is the leading European provider of optimization software. Dash is very interested in this project and will supply test problems to evaluate the potential of the proposed decomposition method.

Campo scientifico (EuroSciVoc)

Programma(i)

Argomento(i)

Invito a presentare proposte

Dati non disponibili

Meccanismo di finanziamento

RGI - Research grants (individual fellowships)

Coordinatore

Technische Universiteit Delft
Contributo UE
Nessun dato
Indirizzo
132,Julianalaan
2600 AJ Delft
Paesi Bassi
Costo totale
Nessun dato

Partecipanti (1)