International Journal of Engineering Trends and Technology (IJETT) – Volume 27 Number 1- September 2015
Energy-Efficient Full Area Coverage using Ant Colony
Optimization Algorithm
Reshmi R1
1
M.Tech scholar, Dept. of ECE, Sree Buddha College of Engineering for Women,
Pathanamthitta, India
Abstract: Network coverage is one of the most critical
issues to implement Wireless Sensor Networks
(WSNs). It is important to find out a sensor set with
maximal residual energy to cover all points in the
region of interest. This issue was named as minimum
weight ε-full area coverage problem (MWFCP).
Existing work uses ε-FCO algorithm to solve this
problem. Here in the proposed work, an intelligent
algorithm called ant colony optimization algorithm
(ACO) is used to solve this problem, yielding a better
solution to MWFCP to extend network lifetime. The
ability of ants in finding the shortest path between
their nest (source node) and the food source (data
sink) is the inspiration for ACO algorithm. Simulation
results are conducted to demonstrate the effectiveness
of proposed algorithm over existing algorithm.
Keywords- Wireless Sensor Network, Minimum Weight εfull area coverage problem,
Optimization (ACO) algorithm.
ε-FCO,
Ant
Colony
I. INTRODUCTION
Wireless Sensor Network (WSN) consists of a large
number of sensor nodes, which are used to monitor the
environmental changes in the region of interest (ROI).
Due to this monitoring capability, sensor networks
have a wide range of applications such as intrusion
detection, military applications, target tracking,
environmental monitoring, etc. All these applications
put a lot of constraints on the WSNs. These
constraints create certain number of issues in the
network [2].Among them, coverage is considered as
the fundamental issue in WSN.
Coverage means how well the deployed sensors are
able to monitor each of the event happening points in
the ROI [1]. Usually coverage is associated with two
WSN properties; connectivity and network lifetime.
Connectivity creates a path between the data source
and data sink. For successful data transmission
connectivity is a must.
Once the sensor node is deployed, it is difficult to
frequently change the battery when it is out of charge.
So we have to design an energy efficient sensor
network.
Various algorithms have been proposed to solve the
area coverage problem. This thesis focus on the
following considerations: evaluating and improving
area coverage, while maintaining connectivity and
maximizing the network lifetime.
ISSN: 2231-5381
II. RELATED WORKS
Qianqian Yang et al. [1] proposed a system based
on probabilistic sensing model to provide full area
coverage in wireless sensor networks (WSNs). They
describe the probabilistic sensing model as a more
realistic model for characterizing the sensing region.
This work gives a brief idea about theminimum
weight ε-full area coverage problem (MWFCP) and
the solution ε-FCO algorithm.
Raymond Mulligan and Habib M Ammari
conducted a survey on coverage problems in WSNs.
Based on their findings connectivity is closely related
to network lifetime. By maintaining connectivity
between sensor nodes and the data sink the lifetime of
WSNs can increased [3]. Various factors that affect
the coverage performance of WSNs such as various
deployment strategies, node types [4].
Jiming Chen et al. [5] introduce various intelligent
algorithms to solve the coverage problems in sensor
networks. Among all these intelligent algorithms ant
colony optimization algorithm (ACO) gives the better
results. R. Seidlova and J. Pozivil use such algorithms
in combinational optimization tasks [6].Biological
analogy of ACO were discussed in [10]. The main
sections in ACO were discussed in [7]-[10].
III. PROBLEM FORMULATION AND SOLUTION
A. Probabilistic sensing model
The probabilistic sensing model in which a sensor
is able to sense a target at a distance d away with
probability λ(d), whereλ(d) is a decreasing function,
i.e., 0 ≤λ(d) ≤ 1. Under the probabilistic sensing model,
the coverage is defined as the probability of point in
the ROI being covered by the network. The network
can provide ε-full coverage if every point can be
covered with a probability larger than ε, here ε, 0 <ε<
1, is an application dependent threshold [1].
B. Minimum Weight ε-Full Area Coverage Problem
(MWFCP)
Given ROI R, a set of N sensors, the weight
coefficient ωi associated with each sensor Si, i= 1,
2, . . . , N , the Minimum Weight ε-Full Area
Coverage Problem (MWFCP) is to find a subset of
sensors with the minimum aggregate weight to
provide ε-full area coverage to ROI R. The weight
coefficient is a parameter that is negatively correlated
to the residual energy of sensor, i.e., a sensor with
more residual energy has lower weight coefficient.
Activating a set with minimum aggregate weight
http://www.ijettjournal.org
Page 11
International Journal of Engineering Trends and Technology (IJETT) – Volume 27 Number 1- September 2015
during each time slot can efficiently reduce energy
and balance the energy consumption among deployed
sensors to improve the network lifetime.
The MWFCP is divided into two sub problems.
1) Find a set of anchor points (APoint) that are set
PS in ROI such that any point in R is within the
coverage range of at least one point in APoint.
2) Find the minimum weight cover such that each
point Pj in PS can be detected at a probability no less
than εj , whereεj=εekrj, and rj is the ε-coverage range of
Pj. Hence, the coverage of an arbitrary point p in R is
guaranteed to be at least ε.
In MWFCP, a subset of sensors should be selected
to minimize the total weight and guarantee ε-full
coverage. MWFCP can be formulated as an integer
programming problem [1]. Here ACO is developed to
solve this MWFCP.
Initialize number of ants;
Initialize the ACO parameters;
while not end-condition do
for k=0 to number of ants
ant k starts from a random node;
while solution is not constructed do
ant k
probability node;
selects
higher
end while
end for
Local search procedure;
Update-pheromone-trails;
C. ACO algorithm
end while
Ant colony optimization (ACO) algorithms have
been successfully applied to combinatorial
optimization tasks. The ACO algorithm is based on
the behavior of ants in searching of food [10].
Social insects like ants, bees and termites work by
themselves in their simple tasks, independently of
others members of the colony. However, when they
act as a community, they are able to solve complex
problems emerging in their daily lives, by means of
mutual co-operation. This emergent behavior of a
group of social insects is known as swarm
intelligence.Ants are able to find the shortest path
between a food source and the nest without the aid of
visual information, and also to adapt to a changing
environment. It was found that the way ants
communicate with each other is based on pheromone
trails. While ants move, they drop a certain amount of
pheromone on the floor, leaving behind a trail of this
substance that can be followed by other ants. The
more ants follow a pheromone trail, the more
attractive the trail becomes to be followed in the near
future [6].The basic idea is illustrated in Fig.1.
In practical cases, ACO algorithm uses a colony of
artificial ants that behave as co-operative agents in a
mathematic space were they are allowed to search and
reinforce pathways (or solutions) in order to find the
optimal ones. The problem is represented by graph
and the ants walk on the graph to construct solutions.
The solution is represented by a path in the graph.
After initialization of the pheromone trails, ants
construct feasible solutions, starting from random
nodes, then the pheromone trails are updated. At each
step ants compute a set of feasible moves and select
the best one (according to some probabilistic rules
based on a heuristic guided function) to carry out the
rest of the tour. The structure of ACO algorithm is
shown below:
ISSN: 2231-5381
Figure 1: (A) Real ants follow a path between nest and food source.
(B) An obstacleappears on the path: Ants choose whether to turn
left or right with equal probability.(C) Pheromone is deposited
more quickly on the shorter path. (D) Allants have chosen the
shorter path.
ACO consists of two main sections: initialization
and a main loop [9]. The main loop runs for a user
defined number of iterations. N-nodes is denoted as,
S= {P1, P2,…,PN};Pi = (xi,yi)
(1)
where a data area is normalized from 0 to 1, and Piis
an i-th node’s position (i= 1, 2,…,N). Each ant (total
M) is deposited on a node selected at random.
Step 1: Initialization:
a. Set initial parameters that are system: variable,
states, function, input, output, input trajectory, output
trajectory.
b. Set initial pheromone trails value.
c. Each ant is individually placed on initial state with
empty memory.
Let an iteration number t = 0. τij(t) is an
amount of pheromone deposited on a path (i; j)
between the node i and the node j at time t, and τij(t) is
initially set to τ0.
http://www.ijettjournal.org
Page 12
International Journal of Engineering Trends and Technology (IJETT) – Volume 27 Number 1- September 2015
Step 2: While termination conditions not meet do
a. Construct Ant Solution (Find tour): Each ant
constructs a path by successively applying the
transition function the probability of moving from
state to state depend on: as the attractiveness of the
move, and the trail level of the move.
For the ants the visiting node is chosen by the
probability pij(t).The probability of k-th ant moving
from the node i to the node j is decided by,
the edges that they traverse. The amount of
pheromone deposited may depend on the quality of
the solution found. Subsequent ants use the
pheromone information as a guide toward promising
regions of the search space. Ants adaptively modify
the way the problem is represented and perceived by
other ants, but they are not adaptive themselves[9].
At each stage, the ant chooses to move from
one node to another according to some rules:
I.
It must visit each node exactly once.
II.
A distant node has less chance of
being chosen (the visibility).
(2)
III.
The more intense the pheromone
where k = 1,2,.., M, and 1/ηij is a distance of the path(i;
trail laid out on anedge between two
j). The adjustable parameter β control the weight of
nodes, the greater the probability
the node information of the ant. Therefore, the
that that edge will be chosen.
searching ability goes up and down by changing α and
IV.
Having completed its journey, the
β. Ants judge next node by the pheromone and the
ant deposits more pheromones on all
distance from a present location. Nk is a set of noes
edges it traversed, if the journey is
that k-th ant has never visited. The ants repeat
short.
choosing next node until all the nodes are visited.
V.
After each iteration, trails of
b. Apply Local Search
pheromones evaporate [8].
c. Best Tour check: If there is an improvement, update
it.
IV. PERFORMANCE EVALUATION
d. Update Trails:
Simulation results are conducted to demonstrate the
- Evaporate a fixed proportion of the
effectiveness
of our proposed algorithm in terms of
pheromone on each path.
network
lifetime
over existing algorithms.
- For each ant perform the “ant-cycle”
pheromone update.
- Reinforce the best tour with a set number of
“elitist ants” performing the “ant-cycle”.
After all ants have completed their tours, the
amount of deposited pheromone on each path is
updated. Then, the tour length Lk(t) is computed for
each ants, and the amount of pheromone Δτkij(t)
deposited by k-th ant on the path (i; j)is decided as,
(3)
where Tk(t) is a tour obtained by k-th ant, and Lk(t) is
its length. Update τij(t) of each path (i; j) depending on
its Δτkij(t),
Figure.2: Average residual energy of sensors in active sensor cover
(4)
where ρ=[0; 1] is the rate of pheromone evaporation.
e. Create a new population by applying the following
operation, based on pheromone trails. The operations
are applied to individual(s) selected from the
population with a probability based on fitness.
Let t = t + 1. Go back to [Step2] and repeat until t =
tmax[7].
End While
When an ant complete solution, or during the
construction phase, the ant evaluates the solution and
modifies the trail value on the components used in its
solution. Ants deposit a certain amount of pheromone
on the components; that is, either on the vertices or on
ISSN: 2231-5381
Performance of network lifetime is focused
in the simulations. The operation time is divided into
time slots. At initialization stage of each slot,
algorithms are performed to select a set of sensors
with a minimum aggregated weight. The selected
sensors are active and the others are turned off in the
current slot. Assume that a sensor does not consume
any energy in its inactive state. Each sensor node has
equal initial energy, which can last for ten active slots.
As simulation runs, the weight associated with each
sensor increases in accordance with its energy
consumption. Simulation ends when the network fails
to cover any point in the monitored region.
Simulations for the ε-FCO as well as ACO
proposed in this paper under the same experimental
setting were performed here.
http://www.ijettjournal.org
Page 13
International Journal of Engineering Trends and Technology (IJETT) – Volume 27 Number 1- September 2015
The average residual energy of sensor nodes
in the selected set is an important metric to evaluate
the performance of algorithms and thus the results
were shown in Fig. 2. We can see from Figs. 2 and 3
that ACO generally select sensors with more residual
energy than the ε-FCO algorithm, leading to a balance
of energy consumption among all sensors. The total
numbers of active sensors, which is used to
characterize energy consumption of sensor network
per slot, is plotted in Fig. 3. Due to the efficiency of
the applied intelligent algorithm, ACO, generally
activate sensors with more residual energy and need
less number of sensors to cover the all points in the
given ROI.
[6]
[7]
[8]
[9]
[10]
R. Seidlov and J. Poivil, ”Implementation of Ant Colony
Algorithms in MATLAB”
Sho Shimomuray, Masaki Sugimotoy, Taku Haraguchiy,
Haruna Matsushitaz and Yoshifumi Nishioy,” Ant Colony
Optimization with Intelligent and Dull Ants”, 2010
International Symposium on Nonlinear Theory and its
Applications, NOLTA2010, Krakow, Poland, September 5-8,
2010
Anjali1, Savita and Gurpal Singh,”Comparing Ant Colony
Optimization & Genetic Algorithm for Solving Energy
Efficient Coverage in WSNs”, International Journal For
Advance Research In Engineering And Technology, Vol. 1,
Issue II, Mar. 2013
Nada M. A. Al Salami,“Ant Colony Optimization
Algorithm”, UbiCC Journal, Vol. 4, No. 3, Aug. 2009.
Marco Dorigo and Thomas St¨utzle,“The Ant Colony
Optimization Metaheuristic:Algorithms, Applications, and
Advances”,2011
Figure 3: Number of active sensors in each time slot
V. CONCLUSION
In this dissertation, area coverage with the
optimization goal of maximizing network lifetime by
designing ACO algorithm has been focused. This
algorithm develops a more sophisticated heuristic to
improve the network lifetime. The problem of
maintaining both area coverage and network
connectivity under energy constraint in WSN has been
extensively addressed here. Along with the solutions,
the theoretical analysis and extensive simulations
prove the efficiency of proposed work.
REFERENCES
[1]
[2]
[3]
[4]
[5]
Qianqian Yang, Shibo He, Junkun Li, Jiming Chen and
Youxian Sun, ”Energy-Efficient Probabilistic Area Coverage
in Wireless Sensor Networks”, IEEE Transactions on
Vehicular Technology, vol. 64, no. 2, Jan. 2015.
Jyoti Yadav and Sandeep Mann, ”Coverage in Wireless
Sensor Networks: A Survey”, International Journal of
Electronics and Computer Science Engineering, ISSN 22771956, vol. 2, no. 1, Jan. 2015, pp.465-471.
Raymond Mulligan and Habib M Ammari, ”Coverage in
Wireless Sensor Networks: A Survey”, Network Protocols
and Algorithms, ISSN 1943-3581, 2010, vol. 2, no. 2, pp. 2753.
GaoJun Fan and ShiYao Jin, ”Coverage Problem in Wireless
Sensor Network: A Survey”, Journal of Networks, vol. 5, no.
9, sept. 2010, pp. 1033-1040.
Jiming Chen, Junkun Li, Shibo He, Youxian Sun and HsiaoHwa Chen, ”Energy-Efficient Coverage Based on
Probabilistic Sensing Model in Wireless Sensor Networks”,
IEEE Communication Letters, vol. 14, no. 9, sept. 2010,
pp.833-835
ISSN: 2231-5381
http://www.ijettjournal.org
Page 14