Parallel PIPS-SBB: multi-level parallelism for stochastic mixed-integer programs
Abstract
PIPS-SBB is a distributed-memory parallel solver with a scalable data distribution paradigm. It is designed to solve MIPs with a dual-block angular structure, which is characteristic of deterministic-equivalent Stochastic Mixed-Integer Programs (SMIPs). In this paper, we present two different parallelizations of Branch & Bound (B&B), implementing both as extensions of PIPS-SBB, thus adding an additional layer of parallelism. In the first of the proposed frameworks, PIPS-PSBB, the coordination and load-balancing of the different optimization workers is done in a decentralized fashion. This new framework is designed to ensure all available cores are processing the most promising parts of the B&B tree. The second, ug[PIPS-SBB,MPI], is a parallel implementation using the Ubiquity Generator (UG), a universal framework for parallelizing B&B tree search that has been sucessfully applied to other MIP solvers. We show the effects of leveraging multiple levels of parallelism in potentially improving scaling performance beyond thousands of cores.
- Authors:
-
- Georgia Inst. of Technology, Atlanta, GA (United States)
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Zuse Inst. Berlin (Germany)
- Publication Date:
- Research Org.:
- Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
- Sponsoring Org.:
- USDOE National Nuclear Security Administration (NNSA)
- OSTI Identifier:
- 1635781
- Report Number(s):
- LLNL-JRNL-739981
Journal ID: ISSN 0926-6003; 893506
- Grant/Contract Number:
- AC52-07NA27344
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Computational Optimization and Applications
- Additional Journal Information:
- Journal Volume: 73; Journal Issue: 2; Journal ID: ISSN 0926-6003
- Publisher:
- Springer
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 97 MATHEMATICS AND COMPUTING
Citation Formats
Munguía, Lluís-Miquel, Oxberry, Geoffrey, Rajan, Deepak, and Shinano, Yuji. Parallel PIPS-SBB: multi-level parallelism for stochastic mixed-integer programs. United States: N. p., 2019.
Web. doi:10.1007/s10589-019-00074-0.
Munguía, Lluís-Miquel, Oxberry, Geoffrey, Rajan, Deepak, & Shinano, Yuji. Parallel PIPS-SBB: multi-level parallelism for stochastic mixed-integer programs. United States. https://doi.org/10.1007/s10589-019-00074-0
Munguía, Lluís-Miquel, Oxberry, Geoffrey, Rajan, Deepak, and Shinano, Yuji. Fri .
"Parallel PIPS-SBB: multi-level parallelism for stochastic mixed-integer programs". United States. https://doi.org/10.1007/s10589-019-00074-0. https://www.osti.gov/servlets/purl/1635781.
@article{osti_1635781,
title = {Parallel PIPS-SBB: multi-level parallelism for stochastic mixed-integer programs},
author = {Munguía, Lluís-Miquel and Oxberry, Geoffrey and Rajan, Deepak and Shinano, Yuji},
abstractNote = {PIPS-SBB is a distributed-memory parallel solver with a scalable data distribution paradigm. It is designed to solve MIPs with a dual-block angular structure, which is characteristic of deterministic-equivalent Stochastic Mixed-Integer Programs (SMIPs). In this paper, we present two different parallelizations of Branch & Bound (B&B), implementing both as extensions of PIPS-SBB, thus adding an additional layer of parallelism. In the first of the proposed frameworks, PIPS-PSBB, the coordination and load-balancing of the different optimization workers is done in a decentralized fashion. This new framework is designed to ensure all available cores are processing the most promising parts of the B&B tree. The second, ug[PIPS-SBB,MPI], is a parallel implementation using the Ubiquity Generator (UG), a universal framework for parallelizing B&B tree search that has been sucessfully applied to other MIP solvers. We show the effects of leveraging multiple levels of parallelism in potentially improving scaling performance beyond thousands of cores.},
doi = {10.1007/s10589-019-00074-0},
journal = {Computational Optimization and Applications},
number = 2,
volume = 73,
place = {United States},
year = {Fri Feb 15 00:00:00 EST 2019},
month = {Fri Feb 15 00:00:00 EST 2019}
}
Web of Science
Figures / Tables:
Works referenced in this record:
Branching rules revisited
journal, January 2005
- Achterberg, Tobias; Koch, Thorsten; Martin, Alexander
- Operations Research Letters, Vol. 33, Issue 1
Automatic Dantzig–Wolfe reformulation of mixed integer programs
journal, March 2014
- Bergner, Martin; Caprara, Alberto; Ceselli, Alberto
- Mathematical Programming, Vol. 149, Issue 1-2
PEBBL: an object-oriented framework for scalable parallel branch and bound
journal, July 2015
- Eckstein, Jonathan; Hart, William E.; Phillips, Cynthia A.
- Mathematical Programming Computation, Vol. 7, Issue 4
Progress in presolving for mixed integer programming
journal, June 2015
- Gamrath, Gerald; Koch, Thorsten; Martin, Alexander
- Mathematical Programming Computation, Vol. 7, Issue 4
A high-performance, portable implementation of the MPI message passing interface standard
journal, September 1996
- Gropp, William; Lusk, Ewing; Doss, Nathan
- Parallel Computing, Vol. 22, Issue 6
Could we use a million cores to solve an integer program?
journal, May 2012
- Koch, Thorsten; Ralphs, Ted; Shinano, Yuji
- Mathematical Methods of Operations Research, Vol. 76, Issue 1
An Automatic Method of Solving Discrete Programming Problems
journal, July 1960
- Land, A. H.; Doig, A. G.
- Econometrica, Vol. 28, Issue 3
A Computational Study of Search Strategies for Mixed Integer Programming
journal, May 1999
- Linderoth, J. T.; Savelsbergh, M. W. P.
- INFORMS Journal on Computing, Vol. 11, Issue 2
Parallel distributed-memory simplex for large-scale stochastic LP problems
journal, February 2013
- Lubin, Miles; Hall, J. A. Julian; Petra, Cosmin G.
- Computational Optimization and Applications, Vol. 55, Issue 3
On parallelizing dual decomposition in stochastic integer programming
journal, May 2013
- Lubin, Miles; Martin, Kipp; Petra, Cosmin G.
- Operations Research Letters, Vol. 41, Issue 3
Cutting planes in integer and mixed integer programming
journal, November 2002
- Marchand, Hugues; Martin, Alexander; Weismantel, Robert
- Discrete Applied Mathematics, Vol. 123, Issue 1-3
PIPS-SBB: A Parallel Distributed-Memory Branch-and-Bound Algorithm for Stochastic Mixed-Integer Programs
conference, May 2016
- Munguia, Lluis-Miquel; Oxberry, Geoffrey; Rajan, Deepak
- 2016 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)
The Million-Variable “March” for Stochastic Combinatorial Optimization
journal, July 2005
- Ntaimo, Lewis; Sen, Suvrajeet
- Journal of Global Optimization, Vol. 32, Issue 3
A stochastic programming approach for supply chain network design under uncertainty
journal, November 2005
- Santoso, Tjendera; Ahmed, Shabbir; Goetschalckx, Marc
- European Journal of Operational Research, Vol. 167, Issue 1
Preprocessing and Probing Techniques for Mixed Integer Programming Problems
journal, November 1994
- Savelsbergh, M. W. P.
- ORSA Journal on Computing, Vol. 6, Issue 4
Solving Open MIP Instances with ParaSCIP on Supercomputers Using up to 80,000 Cores
conference, May 2016
- Shinano, Yuji; Achterberg, Tobias; Berthold, Timo
- 2016 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
FiberSCIP—A Shared Memory Parallelization of SCIP
journal, January 2018
- Shinano, Yuji; Heinz, Stefan; Vigerske, Stefan
- INFORMS Journal on Computing, Vol. 30, Issue 1
A two-stage stochastic mixed-integer programming approach to the index tracking problem
journal, November 2009
- Stoyan, Stephen J.; Kwon, Roy H.
- Optimization and Engineering, Vol. 11, Issue 2
Optimization of Collective Communication Operations in MPICH
journal, February 2005
- Thakur, Rajeev; Rabenseifner, Rolf; Gropp, William
- The International Journal of High Performance Computing Applications, Vol. 19, Issue 1
PySP: modeling and solving stochastic programs in Python
journal, March 2012
- Watson, Jean-Paul; Woodruff, David L.; Hart, William E.
- Mathematical Programming Computation, Vol. 4, Issue 2
Faster MIP solutions via new node selection rules
journal, September 2010
- Wojtaszek, Daniel T.; Chinneck, John W.
- Computers & Operations Research, Vol. 37, Issue 9
Computational Experience with a Software Framework for Parallel Integer Programming
journal, August 2009
- Xu, Y.; Ralphs, T. K.; Ladányi, L.
- INFORMS Journal on Computing, Vol. 21, Issue 3
Stochastic Optimization for Unit Commitment—A Review
journal, July 2015
- Zheng, Qipeng P.; Wang, Jianhui; Liu, Andrew L.
- IEEE Transactions on Power Systems, Vol. 30, Issue 4
The million-variable "march" for stochastic combinatorial optimization
text, January 2004
- Ntaimo, Lewis; Sen, Suvrajeet
- Humboldt-Universität zu Berlin, Mathematisch-Naturwissenschaftliche Fakultät II, Institut für Mathematik
Figures / Tables found in this record: