Adaptive Social Hierarchies: From Nature to Networks

by Markham, A.
Abstract:
Social hierarchies are a common feature of the Animal Kingdom and control access to resources according to the fitness of the individual. We use a similar concept to form an Adaptive Social Hierarchy amongst nodes in a heterogeneous wireless network so that they can discover their role in terms of their base attributes (such as energy or connectivity). Three different methods of forming the hierarchy are presented (pairwise, one-way and agent based). With Agent ASH we show that the time taken for the hierarchy to converge decreases with increasing N, leading to good scalability. The ranked attributes are used as a network underlay to enhance the behaviour of existing routing protocols. We also present an example of a cross-layer protocol using ranked connectivity and energy. The ASH hierarchy provides an abstraction of real world, absolute values onto a relative framework, and thus leads to simpler and more general protocol design.
Reference:
Adaptive Social Hierarchies: From Nature to Networks (Markham, A.), Chapter in Bio-Inspired Computing and Networking (Xiao, Yang, ed.), CRC Press, 2011.
Bibtex Entry:
@InCollection{markham2008adaptive,
  Title                    = {Adaptive Social Hierarchies: From Nature to Networks},
  Author                   = {Markham, A.},
  Booktitle                = {Bio-Inspired Computing and Networking},
  Publisher                = {CRC Press},
  Year                     = {2011},
  Editor                   = {Yang Xiao},

  Abstract                 = {Social hierarchies are a common feature of the Animal Kingdom and control access to resources according to the fitness of the individual. We use a similar concept to form an Adaptive Social Hierarchy amongst nodes in a heterogeneous wireless network so that they can discover their role in terms of their base attributes (such as energy or connectivity). Three different methods of forming the hierarchy are presented (pairwise, one-way and agent based). With Agent ASH we show that the time taken for the hierarchy to converge decreases with increasing N, leading to good scalability. The ranked attributes are used as a network underlay to enhance the behaviour of existing routing protocols. We also present an example of a cross-layer protocol using ranked connectivity and energy. The ASH hierarchy provides an abstraction of real world, absolute values onto a relative framework, and thus leads to simpler and more general protocol design.},
  Url                      = {http://www.cs.ox.ac.uk/files/2111/ASH.pdf}
}