Professional Documents
Culture Documents
1
Where:
i
N means that the ant k is allowed to access the
next node collection within communication. ( )
ij
t is the
pheromone strength of ( , )
i j
e v v on t time.
are the
system parameters, which can control the pheromone and
heuristic value of the relative importance. ( )
ij
t is the local
heuristic factors on the ( , )
i j
e v v . ( )
ij
t can be given by
1
2
[ ( )]
( ) ( )
[ ( , )]
j
ij ij
E t
n t t
d i j
= +
?
1
3
2
[ ( )]
( , )
( ) [ ]
[ ( , )] ( , )
j
ij
E t
d j s
t
d i j d i s
=
3
Where: ( )
j
E t is the residual energy of the selected
node
j
v at t time; ( , ) d i j is the distance between the node
i
v and
j
v ; ( , ) d i s and ( , ) d j s are the distance between nodes
i
v and
j
v to
s
v ;
1
2
3
are the weight coefficient.
Formula (1) indicates that the ants find the path is
impacted by the pheromone, the current energy of node, the
distance between each node and the source node .which
making the convergence to the optimal solution,as well as
balanced the energy consumption among nodes, so the
energy consumption is relatively evenly,and effictively
extending the lifetime of the entire network.
B. Pheromone Updata Rules
When the ant k from node
i
v to
j
v through ( , )
i j
e v v ,the
pheromone on ( , )
i j
e v v is modified as the following:
( , 1) (1 ) ( ) ( , 1)
k
ij ij ij
t t t t t + = + + +
2
( , 1) ( ) / ( , )
k
ij i j
t t E E d i j + = +
Where: is the pheromone evaporation coefficient,
0< <1; is the weight coefficient;
( , ) d i j is the distance
between the node
i
v and
j
v ;
i
E
j
E are the residual
energy of the selected node
i
v and
j
v .
III. UNEVEN CLUSTERING ROUTING METHOD
FOR WSNS BASED ON ANT COLONY OPTMIZTION
The uneven clustering routing algorithm for WSNs
based on ant colony optimization has a periodic round; each
round is divided into cluster formation and cluter route stage.
After new round start, the algorithm firstly divided the
cluter, then the data will be transmitted between the CH.
In the clustr formation stage, the base station firstly need
to use a given transmit power to network to broadcast a
signal. After receiving this signal, each sensor according to
the received signal strength to calulate the distance to the
base station. The distance not only helps sensor nodes to
select the appropriate transmit power to transmit data to
base station to conserve energy consumption, but also is one
of the necessary information for sensor nodes to construct
clusters.
Uneven clustering mechanism is introduced when
construct the scale of the cluster. The cluster near the base
station will have fewer cluster members so that it can set
aside part of the energy for CH to forward the data; the
cluster away from the base station will have more
members,because the long-distance communication
comsume more energy. Therefore, as the distance between
the CH and the base station decreases, its competition radius
should be decreased.
Because the CH has to perform data integration and
tansfer information, they consume more energy than the
b8
members. Therefore, after generating the candidate CH,
introducing the competitive algorithm, when each candidate
CH compete with the other candidate CH within the radius
R
c
to select the candidate CH of largest residual energy to
become the final CH.
The main function of cluster route stage is to transmit
data. In order to achieve multi-hop communication between
the clusters, and reduce the energy consumption of the CH
which away from the base station, using the ant colony
optimization algorithm to find the optimal path between the
CH. CH through the optimal path to transmit data to reduce
the energy consumption, and extend the network life cycle.
The algorithm flow chart is shown in Fig.1.
Figure 1. Alogrithm Flow Chart
A. Election of Clusterhead
In the CH election process, the sensor node firstly
generates a random number bwteen 0 and 1, and if less than
the threshold T (n), then this node becomes the candidate CH.
In each round of the loop, if the node is elected as the CH,
put T (n) set to 0, so that the node will not be re-elected as
CH. We modify the equation of T (n) as follows:
0
0,
,
1 [ mod(1/ )]
( )
residual
n Other
E p
n G
p r p E
T n
b
B. The Formation of Cluster
Once the node through the above calculation to become
the CH, according to the distane d which is calculated in
the cluster formation stage to calculate the size of the cluster
radius
c
R . The equation of
c
R is given by
max
0
max min
( , )
(1 )
i
c
d d s DS
R c R
d d
Where:
c is to control the range of parameters, the
range between 0 and 1;
max
d
and
min
d
are the network nodes
to the base stations maximum and minimum distance;
( , )
i
d s DS
is the distance between node
i
s to the base station;
0
R is the maximum range of the cluster. According to this
formula, the closer to base station,
c
R is smaller.
Once
c
R is chosen, each candidate CH broadcasts
competitive news to nodes within its clusters, if there is no
other candidate CH, the node announces as the CH , and
broadcast competitive success message. If there is other
candidate CH within its clusters, then compare the residual
energy with other nodes, the nodes that has the largest
residual energy becomes the final CH and broadcast success
news.
After the CH is identified, other nodes according to the
received signal strength to determine subordinate cluster,
and inform the appropriate CH. When the CH receives the
news, it uses the TDMA method to assigne time slot for the
nodes to send data, and then the cluster is completed.
C. Routing Design of inter-cluster based on ant colony
optimiztion
1) Algorithm Rlue
a) Ant carrys information including the current hop
count and current crawling line
b) Ant search the next hop node by the probability
calculation formula(1)(2)(3)
c) When ant moves form node
i
v to
j
v , according
to the formula (4)(5) to update pheromone
d) If ant finds the convergence node in the hop node,
the ant out of the search
e) When ant moves to the non-destination node, and
can not select the next hop, then exit the search,
the search will no longer be on the path to that
node.
f) The stop condition for ants to search is that the
sink node is found or exceed the limit number of
hops. The final condidtion for ants to stop is that
the length of the crawling line is shortest route.
2) Algorithm Design
Step 1 state initialization: placing K ants to each CH,
initialize the number of hops 1 JNum= ,the number of
iteration max R C = (constant), set and initialize three
matrices
Tabu
_ R best
and
_ A city
,
Tabu
is used to store and
record the generated path,
_ R best
is used to store the best
path for CH,
_ A city
is uesd to store the visited node, the
pheronmone matrix Tau is initialized to matix ( , ) n n .
b9
Step 2 Insert the visited nodes into
_ A city
, each ant
searches the next hop by the rule 2, and updates
_ A city
and
Tabu
.
Step 3 updating the pheromone values on the path by
rule 3
Step 4 determine each ant wether to meet the iteration
termination condition by rule 4,5,6.If the ant does not meet
the conditions, then return to step 2 to continue the search
algorithm, such as to meet the conditions ,then save the
current optimal solution.
IV. SIMULATION RESULTS
A. Energy Model and Paramenters
In this paper, we use the same energy model with
paper[3], and wireless transmitter module can control the
size of the transmit power, according to the distance
between the nodes. We constructed an event driven
simulator by using Matlab 7.0. The specific simulation
environment: In the network, 200 sensor nodes are
randomly distributed on a MM region with M=200m. the
coordinate of the base station is (100,350),
PacketLength
=4000,
ctrPacketLength
=100,
s
R
=20 m,
DA
E =0.5nJ/bit,
elec
E
=50 nJ/bit,
fs
=10 pJ/(bit
2
m ),
mp
=0.0013 pJ/( bit
4
m ),
0
d =75m, c =0.5
0
E =0.5J
0
R =30m, =2, =2,
1
=2,
2
=1,
3
=1, =0.2
B. Analysis of Simulation
Figure 2 presentd the comparison of the simulation for the
average energy consumption of the node with the rounds
between LEACH, PAPA and our algorithm. From Fig.2, we
can see that the total energy of 200 nodes are 100J, the
nodes are all killed at 1150 rounds in LEACH, which
consumes energy is 100J, PARA consume energy is 80J, our
algorithm is 55J, which reduce 45J, so that performance of
the network has been greatly improved, and the energy
consumption is more evenly distributed to all nodes.
Figure 2. Averge Energy Consumption
Figure3 presented the results of the simulatin for the
surivival rate of the nodes wth the rounds. From Fig.3, we
can see that the surivial rate of PARA significantly
improved compared to LEACH, while showing a lower
survival rate of the nodes relatively compared with our
algorithm. The reason why this happens is that LEACH
does not consider the residual energy of the nodes when
selecting the CH and CH directly communicate with base
station, and then excessively consume the energy of the CH.
Although PARA used the ant colony optimization algorithm
to transmit data in multi-hop, it does not consider the
balanced energy consumption of the whole network. So that
the CH near the base station will transmit large amount of
data, and will be premature death.
Figure 3. The Survival Rate of Node
V. CONCLUSION
Ant colony optimization algorithm is applied to routing
mechanism of the inter-clusters in wireless sensor newtorks,
we have proposed a novel uneven clustering routing algorithm
for Wireless Sensor Network based on ant colony optimization.
The algorithm uses the uneven mechanism to make the scale
of the clusters which near the base station smaller than the
nodes which away from the base station. Ant colony
optimization algorithm is used to optimize the path on that
the CH to transmit the data. The simulation results show that
compared with LEACH and PARA, our algorithm has
significantly improved in average energy consumption and
survival rate, and extended the network lift cycle.
ACKNOWLEDGMENT
The work presented in this paper was supported in part
by the Chongqing scientific and technological project (CSTC,
2007BA2003)
REFERENCES
[1] ESTRIN D. Wireless sensor network network tuorial part IV: Sensor
network protocols//Proceedings of the ACM Mobile Computing and
Networking(MobiCom), Atlanta, GA, 2002.
[2] AKTILDIZ I F, SU W, SANKARASUBRAMAN Y, CAYIRCI E, A
Survey on Sensor Network[J]. IEEE Communication Magazine, 2002,
40(8):102-114.
0
[3] HEINZALMAN W R, KULIK J, BALAKRISHMAN H. Adaptive
Protocols for Informatrion Dissemination in Wireless Sensor
Network[C]//proc. of the 5
th
Annual Internation Conference on
Mobile Computing and Networking. New York, NY, USA:ACM
press, 2001:174-185.
[4] YOUNIS O, FAHMY S. A Hybrid Energy Efficent Distributed
Clustering Approach for Ad_Hoc sensor Network[J]. IEEE
Transcation on Mobile Computing. 2004,3(4):660-668.
[5] LI W en, LIU Yang-ping. TONG Diao-sheng. Distributed Data
Collection Rote Algorithm Based on Ant Algorithm in Sensor
Network[J]. Computer Systems, 2005,26(5)788-792.
[6] YE Gang-qiang, LIANG Yan, SUN Shi-yu. Ant Colony-based
Wireless Sensor Network Routing Algorithm[J]. Application Reserch
of Computers, 2008,25(3):715-717.
[7] WANG Qing-zheng, MIN Lin, GUO Zheng-wei, Engergy Blance
Routing Algorithm Based on Ant Colony Optimization in Wireless
Sensor Networks[J]. Application Research of Computers,
2009,26(12):716-723.
[8] BONABEAU E, DORIGO M, THERAULA C. Inspiration for
Optimiztion from Social Insect Behavior[J], Nature, July
2000,406:39-42.
[9] JIANG Yang, WEI Zhao, MAI Xu. A Multipath Routing Protocol
Based on Clustering and Ant Colony Optimiztion for Wireless Sensor
Network[J]. Computer Network and Information Security, 2009,
1:50-60.
[10] LI Cheng-fa, CHEN Gui-hai, YE MAO. An Uneven Cluster-Based
Routing Protocol for Wireless Sensor Network[J]. Chinse Journal of
Computers, 2007,30(1):27-36.
[11] CAMILO T, CARRETO C, SILVA J S, et al. An Energy-efficient
Ant-based Routing Alforithm for Wireless Sensor Networks[C]//proc.
of the International Work shop on Ant Colony Optimization and
Swarm Intelligence. Brussels, Belgium: Springer-Verlag, 2006:49-59.
[12] BANDYOPADHYAY S, COYLEE J. A energy efficient hierachical
clustering algorithm for wireless sensor networks[A]. Proceeding of
the INFOCOM[C]. New York, 2003. 1713-1723.
[13] M.DORIGO, L,M, Gambardella, Ant Colony ststem: Acooperative
learning approach to the traveling saleman problem, IEEE Tranc.
Evol Computer, Vol.1, pp. 53-66,1997.
!