MIT Graduate Invents Algorithm For More Competent Ad-Hoc Networks
MIT graduate student Bernhard Haeupler has ensured with his new algorithm that ad-hoc networks could guarantee the transfer of information across all its nodes. Ad-hoc networks are decentralized networks with no predefined structure and hence communication within the network has not been reliable. Haeupler’s algorithm has proven to be faster than previous algorithms and even won one of two best-student-paper prizes for a new algorithm that aims to resolve the issue.
Bernhard's approach was deterministic as opposed to the previous algorithms that were probabilistic and hence always held the chance of an unsuccessful delivery. Researchers have been looking for distributed algorithms such that all nodes perform the same task to derive a desired global behavior. Haeupler's algorithm works in a series of rounds, where each round is the period in which two adjacent nodes exchange information. In the first round, a node exchanges information with one neighbor, then it selects a second node, and it exchanges information with both of the nodes for two rounds and so on.
[caption id="attachment_45168" align="aligncenter" width="600"]#-Link-Snipped-# Haeupler's Algorithm[/caption]
Source & Image: #-Link-Snipped-#
Bernhard's approach was deterministic as opposed to the previous algorithms that were probabilistic and hence always held the chance of an unsuccessful delivery. Researchers have been looking for distributed algorithms such that all nodes perform the same task to derive a desired global behavior. Haeupler's algorithm works in a series of rounds, where each round is the period in which two adjacent nodes exchange information. In the first round, a node exchanges information with one neighbor, then it selects a second node, and it exchanges information with both of the nodes for two rounds and so on.
[caption id="attachment_45168" align="aligncenter" width="600"]#-Link-Snipped-# Haeupler's Algorithm[/caption]
Source & Image: #-Link-Snipped-#
0