• DocumentCode
    625931
  • Title

    Multiple unicasts, graph guessing games, and non-Shannon inequalities

  • Author

    Baber, Rahil ; Christofides, Damianos ; Dang, Anh N. ; Riis, Søren ; Vaughan, Emil R.

  • Author_Institution
    Sch. of Electron. Eng. & Comput. Sci., Queen Mary, Univ. of London, London, UK
  • fYear
    2013
  • fDate
    7-9 June 2013
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    Guessing games for directed graphs were introduced by Riis [8] for studying multiple unicast network coding problems. It can be shown that protocols for a multiple unicast network can be directly converted into a strategy for a guessing game. The performance of the optimal strategy for a graph is measured by the guessing number, and this number can be bounded from above using information inequalities. Christofides and Markstrom [4] developed a guessing strategy for undirected graphs based on the fractional clique cover, and they conjectured that this strategy is optimal for undirected graphs. In this paper we disprove this conjecture. We also provide an example of an undirected graph for which non-Shannon inequalities provide a better bound on the guessing number than Shannon inequalities. Finally, we construct a counterexample to a conjecture we raised during our work which we referred to as the Superman conjecture.
  • Keywords
    directed graphs; game theory; network coding; protocols; Superman conjecture; directed graphs; fractional clique cover; graph guessing game strategy; information inequalities; multiple unicasts; nonShannon inequalities; optimal strategy; protocols; unicast network coding problems; Channel coding; Cramer-Rao bounds; Entropy; Games; Random variables; Unicast; Upper bound;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Network Coding (NetCod), 2013 International Symposium on
  • Conference_Location
    Calgary, AB
  • Print_ISBN
    978-1-4799-0821-9
  • Type

    conf

  • DOI
    10.1109/NetCod.2013.6570823
  • Filename
    6570823