Programming and Computer Software

, Volume 41, Issue 5, pp 302–306 | Cite as

Improving quality of graph partitioning using multi-level optimization

  • R. K. PastukhovEmail author
  • A. V. Korshunov
  • D. Yu. Turdakov
  • S. D. Kuznetsov
Image Processing


Graph partitioning is required for solving tasks on graphs that need to be distributed over disks or computers. This problem is well studied, but the majority of the results on this subject are not suitable for processing graphs with billions of nodes on commodity clusters, since they require shared memory or lowlatency messaging. One of the approaches suitable for cluster computing is the balanced label propagation, which is based on the label propagation algorithm. In this work, we show how multi-level optimization can be used to improve quality of the partitioning obtained by means of the balanced label propagation algorithm.


Computer Node Multi Level Initial Partitioning Social Graph Edge Contraction 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Garey, M.R., Johnson, D.S., and Stockmeyer, L., Some simplified NP-complete graph problems, Theoretical Comput. Sci., 1976, vol. 1, no. 3, pp. 237–267.zbMATHMathSciNetCrossRefGoogle Scholar
  2. 2.
    Karypis, G. and Kumar, V., A fast and high quality multilevel scheme for partitioning irregular graphs, SIAM J. Sci. Computing, 1998, vol. 20, no. 1, pp. 359–392.MathSciNetCrossRefGoogle Scholar
  3. 3.
    Ugander, J. and Backstrom, L., Balanced label propagation for partitioning massive graphs, Proc. of the Sixth ACM Int. Conf. on Web Search and Data Mining, WSDM'13 (Rome, 2013), Rome: ACM, 2013, pp. 507–516.CrossRefGoogle Scholar
  4. 4.
    Kernighan, B.W. and Lin, S., An efficient heuristic procedure for partitioning graphs, Bell System Tech. J., 1970, vol. 49, no. 2, pp. 291–307.zbMATHCrossRefGoogle Scholar
  5. 5.
    Raghavan, U., Albert, R., and Kumara, S., Near linear time algorithm to detect community structures in largescale networks, Phys. Rev., E 76, 036106. Published September 11, 2007.Google Scholar
  6. 6.
    Dean, J. and Ghemawat, S., MapReduce: Simplified data processing on large clusters, OSDI'04: Sixth Symp. on Operating System Design and Implementation, San Francisco, 2004, pp. 137–150.Google Scholar
  7. 7.
    Zaharia M. et al., Resilient distributed datasets: A faulttolerant abstraction for in-memory cluster computing, Proc. of the 9th USENIX Conf. on Networked Systems Design and Implementation, NSDI'12 (San Jose, 2012), San Jose: USENIX Association, 2012.Google Scholar
  8. 8.
    Stanford Network Analysis Project, LiveJournal social network. 2006. Scholar
  9. 9.
    Backstrom, L. et al., Group formation in large social networks: Membership, growth, and evolution, Proc. of the 12th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, KDD'06 (Philadelphia, 2006), Philadelphia: ACM, 2006, pp. 44–54.CrossRefGoogle Scholar
  10. 10.
    Sampling Online Social Networks, Facebook Social Graph. 2009.$#$facebook_social_graph¯_mhrw_uni.Google Scholar
  11. 11.
    Gjoka, M. et al., Walking in Facebook: A case study of unbiased sampling of OSNs, Proc. of the 29th Conf. on Information Communications, INFOCOM’10 (San Diego, 2010), San Diego: IEEE, 2010, pp. 2498–2506.Google Scholar
  12. 12.
    Gonzalez, J.E. et al., PowerGraph: Distributed graphparallel computation on natural graphs, Proc. of the 10th USENIX Conf. on Operating Systems Design and Implementation, OSDI'12 (Hollywood, 2012), Hollywood: USENIX Association, 2012, pp. 17–30.Google Scholar

Copyright information

© Pleiades Publishing, Ltd. 2015

Authors and Affiliations

  • R. K. Pastukhov
    • 1
    Email author
  • A. V. Korshunov
    • 1
  • D. Yu. Turdakov
    • 1
  • S. D. Kuznetsov
    • 1
  1. 1.Institute for System ProgrammingRussian Academy of SciencesMoscowRussia

Personalised recommendations