You are here

A GENETIC ALGORITHM FOR NETWORK TRANSPORT PROTOCOL PARAMETER OPTIMIZATION

Download pdf | Full Screen View

Date Issued:
2009
Summary:
The fields of wireless communications and mobile networking are rapidly growing and changing. In a mobile ad hoc network, the variation of link characteristics, and frequent and unexpected changes in topology, decreases the performance of commonly used transport protocols, which assume that packet drops occur only in the event of network congestion. Link failure and higher bit error rates may also induce a sudden increase of packet losses, triggering the activation of congestion avoidance mechanisms that reduce the overall transmission rate. In general, a transport protocol has numerous configuration parameters (e.g. window size, retransmission timeout, etc.) that can be adjusted to compensate for these environment effects. Thus, the challenge is to identify what is the best configuration (i.e. combination of parameter values) for a given scenario. In this work, an optimization approach based on genetic algorithms is used to automatically tune, in real time, the parameter values of a transport protocol so that it can adapt itself to different operating conditions. When compared to an exhaustive search over a reduced problem space, the experimental results show that the proposed algorithm can identify the optimal configuration settings to maximize throughput for end-to-end wireless communication over multiple hops.
Title: A GENETIC ALGORITHM FOR NETWORK TRANSPORT PROTOCOL PARAMETER OPTIMIZATION.
26 views
3 downloads
Name(s): Granados Murillo, Adrian, Author
Type of Resource: text
Date Issued: 2009
Publisher: University of West Florida
Language(s): English
Summary: The fields of wireless communications and mobile networking are rapidly growing and changing. In a mobile ad hoc network, the variation of link characteristics, and frequent and unexpected changes in topology, decreases the performance of commonly used transport protocols, which assume that packet drops occur only in the event of network congestion. Link failure and higher bit error rates may also induce a sudden increase of packet losses, triggering the activation of congestion avoidance mechanisms that reduce the overall transmission rate. In general, a transport protocol has numerous configuration parameters (e.g. window size, retransmission timeout, etc.) that can be adjusted to compensate for these environment effects. Thus, the challenge is to identify what is the best configuration (i.e. combination of parameter values) for a given scenario. In this work, an optimization approach based on genetic algorithms is used to automatically tune, in real time, the parameter values of a transport protocol so that it can adapt itself to different operating conditions. When compared to an exhaustive search over a reduced problem space, the experimental results show that the proposed algorithm can identify the optimal configuration settings to maximize throughput for end-to-end wireless communication over multiple hops.
Identifier: WFE0000176 (IID), uwf:60798 (fedora)
Note(s): M.S.
Department of Computer Science
Masters
Subject(s): Genetic algorithms, optimization, transport, wireless networks, parameter tuning
Persistent Link to This Record: http://purl.flvc.org/uwf/fd/WFE0000176
Restrictions on Access: public
Use and Reproduction: http://rightsstatements.org/vocab/InC-EDU/1.0/
Host Institution: UWF

In Collections