Publication - Exploiting Organization and Learned Knowledge to Reduce the Exploration Cost of Routing

Authors: Zafar, Huzaifa; Lesser, Victor; and Ganesan, Deepak
Title: Exploiting Organization and Learned Knowledge to Reduce the Exploration Cost of Routing
Abstract: Large scale multi-agent systems (MAS) benefit greatly from an overlay organization design that guides agents in determining when to communicate, how often, with whom, with what priority and so on. However, this same organization knowledge is not utilized by general-purpose wireless network routing algorithms normally used to support agent communication.
Here, we look at modifying the QRouting algorithm, in particular the confidence based extension of QRouting called CQRouting, to take advantage of the following set of information; 1) The layout of the organization in terms of roles and message flow, 2) Application knowledge in-terms of expected bandwidth, response time and message priority and 3) Sleep cycle knowledge.
In this paper, we show an improved application-level bandwidth and response time by applying organization knowledge to network-level routing algorithms. This increased bandwidth and response time is especially important in communication-intensive and power-limited application settings such as agent-based sensor networks where node availability and link dynamics make providing sufficient inter-agent communication especially challenging.
Publication: University of Massachusetts Amherst Computer Science Technical Report #07-68
Date: December 2007
Sources: PDF: /Documents/zafar_tr07-68.pdf
Reference: Zafar, Huzaifa; Lesser, Victor; and Ganesan, Deepak. Exploiting Organization and Learned Knowledge to Reduce the Exploration Cost of Routing. University of Massachusetts Amherst Computer Science Technical Report #07-68. December 2007.
bibtex:
@techreport{Zafar-445,
  author    = "Huzaifa Zafar and Victor Lesser and Deepak Ganesan",
  title     = "{Exploiting Organization and Learned Knowledge to
               Reduce the Exploration Cost of Routing}",
  month     = "December",
  year      = "2007",
  url       = "http://mas.cs.umass.edu/paper/445",
}