Imitation as the Simplest Strategy for Cooperation


Durmus Y., Onur E.

IEEE 23rd International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), Sydney, Avustralya, 9 - 12 Eylül 2012, ss.863-869 identifier identifier

  • Cilt numarası:
  • Doi Numarası: 10.1109/pimrc.2012.6362905
  • Basıldığı Şehir: Sydney
  • Basıldığı Ülke: Avustralya
  • Sayfa Sayıları: ss.863-869

Özet

Ad hoc networks comprise independent cooperative nodes which work together to constitute a system having a value greater than the sum of the values of the individual components. The nodes cooperate to gain access to the medium or to establish a messaging infrastructure by relaying foreign packets. However, when nodes in an ad hoc network operate autonomously without a central authority, they tend to defect, e.g., do not forward each other's packets following the game theoretic analysis. External mechanisms may preserve and enforce cooperation in network in return of additional operational costs or security overheads. However, low power devices may lack computational power that is required to implement the system. Recent works in evolutionary game theory have shown that cooperation may survive in a lattice structured biological network without any enforcement. The spatial structure of the network may allow the survival of the cooperative nodes when they imitate the dominant surrounding strategy. Imitating strategy helps low power devices adapt dynamically to the environment rather than giving deterministic and static decisions. In this work, we apply the imitation strategy to ad hoc networks which have geometric random network structure different from the lattice structured networks. Simulations show that simple imitation strategy allows cooperation to be spread over the network.