JOURNAL OF NETWORKS (JNW)
ISSN : 1796-2056
Volume : 3    Issue : 3    Date : March 2008

Adaptive Route Selection Policy Based on Back Propagation Neural Networks
Fang Jing, R.S.Bhuvaneswaran, Yoshiaki Katayama and Naohisa Takahashi
Page(s): 34-41
Full Text:
PDF (465 KB)


Abstract
One of the key issues in the study of multiple route protocols in mobile ad hoc networks (MANETs)
is how to select routes to the packet transmission destination. There are currently two route
selection methods: primary routing policy and load-balancing policy. Many ad hoc routing protocols
are based on primary (fastest or shortest but busiest) routing policy from the self-standpoint of traffic
transmission optimization of each node. Load-balancing protocols equalize transmission load
among multiple routes in the network. However, the lack of global perspective can cause
congestion in primary policy and prolong delay time in load-balancing policy. So, although they are
sometimes efficient, these two types of policies cannot adapt to intricately changing network
conditions. We propose a new multiple route protocol with an Adaptive route selection Policy based
on a Back propagation Neural network (APBN) to optimize selection policy. In our study, we used a
gradient ascent algorithm to determine the relationship between different optimum route selection
polices and varying conditions in the communication network and to make a neural network that
learns this relationship using the Back Propagation (BP) algorithm to predict the next optimum route
selection policy. We evaluated our protocol using Omnet simulator. The results show that the
proposed scheme performs better than current protocols.

Index Terms
mobile ad hoc network, multiple route, back propagation, neural network, gradient ascent algorithm