30-05-2012, 04:35 PM
Bio-inspired routing protocol for mobile ad hoc networks
[attachment=23592]
Introduction
A mobile ad hoc network (MANET) [1] is a group of mobile
nodes which operate themselves through wireless links. In
contrast with the conventional networks, a MANET does
not need an infrastructure, since nodes rely on each other
to operate themselves, forming what is called multihop
communication. Such networks have more problems and
disadvantages than a conventional network.
Related works
MANETs have special characteristics that must be taken into
account when a routing protocol is implemented. There are
many solutions (RFC 3626 [2], RFC 3561 [3], RFC 4728
[4], RFC 3684 [5], etc.). All these protocols have valid
solutions, but they usually have a specific topology and
characteristics of certain scenarios as a design basis. They
are not always particularly suitable if there are drastic
changes in the dynamic topology of the network. There is a
group of algorithms or routing protocols called bioinspired,
whose essential characteristic consists of being
adaptive, which is especially noteworthy in this kind of
network. The concept of swarm intelligence [6] is
specifically referred to in the literature.
AntHocNet
AntHocNet [15, 16] is a hybrid ACO routing algorithm.
Data from 2004 and 2005 have had numerous extensions
and a great impact, being a pioneer algorithm in this field.
This protocol is applied to multipath and dynamic
networks, that is, creating multiple paths to transmit data
from source to destination in the same data session.
AntHocNet follows a structure similar to AntNet-fast ant
(AntNet-FA) [17], but it differs from AntNet-FA given
that topologies of static networks are applied and
convergence is slow. So, what all ants have to do is choose
the path. AntHocNet, meanwhile, takes into account the
dynamic topology and other characteristics of ad hoc
networks. When the network topology changes, then it
must be restored quickly and this is achieved through a new
route discovery process.
Use distance metric in path exploration
Our approach takes into account the number of hops for the
routes which have been found to be the best. There is a hop
limit on the nodes. This hop limit is established according
to previously calculated routes that have a smaller distance
in hop number. Fig. 7 shows an example of how this
mechanism which we have mentioned in our approach works.
Parameter setting
To evaluate the behaviour of the protocol network simulator 3
(NS-3) has been used [20]. We compared AntOR-DLR
(disjoint-link route version) with AntHocNet. So, we have
used common parameters for both the protocols, as is shown
in Table 1. Table 2 shows the internal characteristics of both
protocols that are compared. Finally, in Table 3 we indicate
the parameters of the scenario, where you can see mobility as
an important characteristic: random waypoint (RWP). We
also used a highly dynamic scenario (maximum speed of
nodes is 10 m/s) because it is attempting to analyse the
worst case. Next, we show the obtained results by comparing
both protocols using the same scenario, according to
different performance parameters.
[attachment=23592]
Introduction
A mobile ad hoc network (MANET) [1] is a group of mobile
nodes which operate themselves through wireless links. In
contrast with the conventional networks, a MANET does
not need an infrastructure, since nodes rely on each other
to operate themselves, forming what is called multihop
communication. Such networks have more problems and
disadvantages than a conventional network.
Related works
MANETs have special characteristics that must be taken into
account when a routing protocol is implemented. There are
many solutions (RFC 3626 [2], RFC 3561 [3], RFC 4728
[4], RFC 3684 [5], etc.). All these protocols have valid
solutions, but they usually have a specific topology and
characteristics of certain scenarios as a design basis. They
are not always particularly suitable if there are drastic
changes in the dynamic topology of the network. There is a
group of algorithms or routing protocols called bioinspired,
whose essential characteristic consists of being
adaptive, which is especially noteworthy in this kind of
network. The concept of swarm intelligence [6] is
specifically referred to in the literature.
AntHocNet
AntHocNet [15, 16] is a hybrid ACO routing algorithm.
Data from 2004 and 2005 have had numerous extensions
and a great impact, being a pioneer algorithm in this field.
This protocol is applied to multipath and dynamic
networks, that is, creating multiple paths to transmit data
from source to destination in the same data session.
AntHocNet follows a structure similar to AntNet-fast ant
(AntNet-FA) [17], but it differs from AntNet-FA given
that topologies of static networks are applied and
convergence is slow. So, what all ants have to do is choose
the path. AntHocNet, meanwhile, takes into account the
dynamic topology and other characteristics of ad hoc
networks. When the network topology changes, then it
must be restored quickly and this is achieved through a new
route discovery process.
Use distance metric in path exploration
Our approach takes into account the number of hops for the
routes which have been found to be the best. There is a hop
limit on the nodes. This hop limit is established according
to previously calculated routes that have a smaller distance
in hop number. Fig. 7 shows an example of how this
mechanism which we have mentioned in our approach works.
Parameter setting
To evaluate the behaviour of the protocol network simulator 3
(NS-3) has been used [20]. We compared AntOR-DLR
(disjoint-link route version) with AntHocNet. So, we have
used common parameters for both the protocols, as is shown
in Table 1. Table 2 shows the internal characteristics of both
protocols that are compared. Finally, in Table 3 we indicate
the parameters of the scenario, where you can see mobility as
an important characteristic: random waypoint (RWP). We
also used a highly dynamic scenario (maximum speed of
nodes is 10 m/s) because it is attempting to analyse the
worst case. Next, we show the obtained results by comparing
both protocols using the same scenario, according to
different performance parameters.