Embarrassingly Parallel Search Reengineered

Derval, Guillaume;Schaus, Pierre;et.al.
(2016) Doctoral Program of the 22nd International Conference on Principles and Practice of Constraint Programming (CP 2016) — Location: Toulouse, France (5.September.2016)

Files

cp2016dp-eps.pdf
  • Open Access
  • Adobe PDF
  • 706.63 KB
Download

Details

Authors
  • Derval, Guillaumeorcid-logoUCLouvain
    Author
  • Schaus, Pierreorcid-logoUCLouvain
    Author
  • et. al.
Abstract
With the current cloud trend, parallelizing applications has become more and more important, and the constraint solvers will have to take profit from this massively parallel computing power now available. Embarrassingly Parallel Search is one of the most promising methods, but some of its aspects have to be explored further. This paper presents improvements on the architecture and methods used by EPS, such as decomposition strategies, bound information sharing, and subproblem representation. An open-source implementation of EPS, built on top of OscaR is also presented.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Derval, G., Schaus, P., & et al. (2016). Embarrassingly Parallel Search Reengineered. “Doctoral Program of the 22nd International Conference on Principles and Practice of Constraint Programming (CP 2016)”, Toulouse. Published. Doctoral Program of the 22nd International Conference on Principles and Practice of Constraint Programming (CP 2016), Toulouse, France. https://hdl.handle.net/2078.5/179806 (Original work published 2016)