In this paper we face a very fundamental problem in Operations Research: to find good dual bounds to generic mixed integer mathematical programs (MIPs) as quickly as possible. In particular, we focus on the scenario where large scale data needs to be considered, multicore CPU architectures are available, and massive parallelism can be exploited by means of decomposition methods. We consider column generation techniques to solve extended formulations obtained by means of Dantzig-Wolfe decomposition for MIPs. We propose a concurrent algorithm, that relaxes the synchronized behavior of classical column generation. Our approach relies on simple data structures and efficient synchronization, still providing the same global convergence properties of classical sequential column generation methods. We present and discuss the results of an extensive experimental campaign, comparing our concurrent algorithm to both a naive parallelization of column generation and the cutting planes algorithm implemented in state-of-the-art commercial optimization packages, considering large scale datasets of a hard packing problem from the literature as representative benchmark. Our approach turns out to be on average one order of magnitude faster than competitors, attaining almost linear speedups as the number of available CPU cores increases.

Asynchronous Column Generation / S. Basso, A. Ceselli - In: Workshop on Algorithm Engineering and Experiments : Proceedings / [a cura di] S. Fekete, V. Ramachandran. - [s.l] : Society for Industrial and Applied Mathematics, 2017. - ISBN 9781611974768. - pp. 197-206 (( Intervento presentato al 19. convegno ALENEX tenutosi a Barcelona nel 2017 [10.1137/1.9781611974768.16].

Asynchronous Column Generation

S. Basso;A. Ceselli
2017

Abstract

In this paper we face a very fundamental problem in Operations Research: to find good dual bounds to generic mixed integer mathematical programs (MIPs) as quickly as possible. In particular, we focus on the scenario where large scale data needs to be considered, multicore CPU architectures are available, and massive parallelism can be exploited by means of decomposition methods. We consider column generation techniques to solve extended formulations obtained by means of Dantzig-Wolfe decomposition for MIPs. We propose a concurrent algorithm, that relaxes the synchronized behavior of classical column generation. Our approach relies on simple data structures and efficient synchronization, still providing the same global convergence properties of classical sequential column generation methods. We present and discuss the results of an extensive experimental campaign, comparing our concurrent algorithm to both a naive parallelization of column generation and the cutting planes algorithm implemented in state-of-the-art commercial optimization packages, considering large scale datasets of a hard packing problem from the literature as representative benchmark. Our approach turns out to be on average one order of magnitude faster than competitors, attaining almost linear speedups as the number of available CPU cores increases.
No
English
Settore INF/01 - Informatica
Settore MAT/09 - Ricerca Operativa
Intervento a convegno
Esperti anonimi
Pubblicazione scientifica
   Towards Research on decomposition Methods for Next Generation Analytics
   FONDAZIONE CARIPLO
   2015-0717
Workshop on Algorithm Engineering and Experiments : Proceedings
S. Fekete, V. Ramachandran
Society for Industrial and Applied Mathematics
2017
197
206
10
9781611974768
Volume a diffusione internazionale
ALENEX
Barcelona
2017
19
Convegno internazionale
Intervento inviato
crossref
Aderisco
S. Basso, A. Ceselli
Book Part (author)
reserved
273
Asynchronous Column Generation / S. Basso, A. Ceselli - In: Workshop on Algorithm Engineering and Experiments : Proceedings / [a cura di] S. Fekete, V. Ramachandran. - [s.l] : Society for Industrial and Applied Mathematics, 2017. - ISBN 9781611974768. - pp. 197-206 (( Intervento presentato al 19. convegno ALENEX tenutosi a Barcelona nel 2017 [10.1137/1.9781611974768.16].
info:eu-repo/semantics/bookPart
2
Prodotti della ricerca::03 - Contributo in volume
File in questo prodotto:
File Dimensione Formato  
1%2E9781611974768%2E16.pdf

accesso riservato

Tipologia: Publisher's version/PDF
Dimensione 320.17 kB
Formato Adobe PDF
320.17 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2434/472432
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? ND
social impact