You are on page 1of 6

Hybrid Butterfly Based Particle Swarm Optimization

for Optimization Problems


Aashish Kumar Bohre1

Dr. Ganga Agnihotri2

Dr. Manisha Dubey3


1, 2, 3

Electrical Engineering Department


MANIT, Bhopal, INDIA1, 2, 3
Abstract- One of the superior optimization algorithms amongst all
earlier introduced algorithms is the particle swarm optimization
algorithms. This paper introduces Butterfly Particle swarm
optimization (BF-PSO) with some novel control parameters such
as sensitivity of butterfly towards nectar by different means of
communication and probability of the nectar presence. This new
hybrid algorithm is based on the intelligent characteristics and
behavior of butterfly during the process of food (nectar) search
and mimics their intelligent network structures. Sensitivity and
the probability of nectar, according to the degree of nodes is
calculated using this new algorithm. By adding the effect of these
modifications in the standard Particle Swarm Optimization
(PSO), the algorithm performance and the ability to search
optimum value of the Particle Swarm Optimization is improved.
Finally, the results of applying the BF-PSO on benchmark
functions are shown. The overall improvement in performances
of the BF-PSO on the basis of the sensitivity of butterfly and
probability of nectar source.
Keywords- Particle Swarm Optimization (PSO), BF-PSO
(ButterflyPSO), sensitivity of butterfly (s), probability of nectar (p).

I.

INTRODUCTION

The development of artificial intelligence systems doesn't


necessitate the entire copy of the natural system; it requires
exploration of ideas, model, and behaviour for an
implementation. It is potential (or intelligence) of individual
insect to identify the various complexity of the tasks and play
or perform for them. Some of the best examples include the
forging process, communication of information between
insects, collection and processing (feeding) of food, mating
process between male and female insects and various other
interesting behaviours, which are quite commonly seen in the
natural environment.
Wilson [1] outlined biological communication as a
natural process on the theatrical character of one cell (or
organism) that changes the probability pattern of characters
(or behaviour) in associate to other another cell (or organism)
in the way of adoption.
Tristram D. Wyatt, [2] describes the behaviour of
insects and animals. The importance of chemical
communication is described with some examples from a
varied variety of insects, animals, moreover as humans,
butterfly or moths, marine copepods, Caenorhabditis elegans,
Drosophila, goldfish, snakes, mice and elephants. The

c
978-1-4799-3486-7/14/$31.002014
IEEE

ecology, evolution and behavior, offers an associate degree


introduction to the fast progress in our understanding of
olfaction at the molecular and neurologic level.
Eberhart, R. C., and Kennedy, J. [3-5] reported the
optimization of nonlinear functions; exploitation particle
swarm methodology is delineated.
The intelligence characteristics and behavior of
artificial systems are aging, dying, fight, think, learn,
collaborate and perish. Frequently, when introduce the
different Artificial intelligence researchers' attempts for
illustrates complicated behaviors of artificial systems. This is
gained from relatively simple principles. The evolution of the
Darwins theory and the concept of the survival of the fittest
for the artificial systems were initiated in the earlier decades
(such as genetic algorithms, evolutionary programming and
etc.) [7-11, 21].
In this paper optimization technique based on the butterfly
is reported. The motivation towards the butterfly based swarm
optimization is searching of food processing, intelligence and
behavior. The searching process of butterflies basically
concentrated on the food source that is nectar sources. The
butterflies have the natural sensitivity to sense the nectar
probability. The butterfly develops an interactive intelligent
system with high communication to find the optimal solutions
[12- 19].
II.

THE PARTICLE SWARM OPTIMIZATION


TECHNIQUE

The Particle Swarm optimization comes under the class of


Swarm Intelligence techniques [3-5]. The particle swarm
optimization was suggested in 1995, by Kennedy and Eberhart
[22,23]. Its a popular and more effective optimization
technique based on the population search. The PSO is a
stochastic method of optimizing simulated the intelligent
swarms behaviour to solve the different nonlinear problems
in optimization based on population. The PSO concept
inspired by mimicking the foraging process (social behaviour),
the swarms like as fishes, birds, etc. By putting a random
velocity to each particle, it acts through the search space
corresponding to the target part, limited by problem specific
constraints. Each particle is evaluated for its fitness based
on the output of the objective function. The particle is

172

attracted to the best fitness location attained by it as well as


that attained across the population so far. The each and every
particle follows the particular velocity throughout the
specified region of the problem. The velocity and its direction
updates by the previous best experiences of particles and
memorized, its the best experience by the experiences of
neighboring (social-influence) [20, 22, 23]. Hence the way in
which, the particle has a leaning to fly towards a more optimal
region within the hunting field or space. [3, 4, 5, 9, 11, 20].
The position of particle updates based on the
updating of the velocity of each particle within the current
iteration. In the next iteration the each particle position is
calculated by summing the velocity and current position. Let
us assume that the particles position and velocity
respectively
in
the
dimension
space
d
are
 and . Finally the
each particle velocity is updated by the equations given below
(1) and (2) [3, 4, 5, 20, 22, 23].

v( k +1) = w * vk + c1 r1 ( pbestk currentposition)


+ c2 r2 ( gbestk currentposition)

x ( k +1) = x k + v k +1

(1)

(2)

Where
N: total size of swarms.
t: the iteration number.

takes up. The fourth stage also last phase of butterfly life
cycle. The last phase of life cycle is only the butterfly. The
butterfly comes out of the chrysalis. Immediately they can
learn to fly, and also can get a mate. When they find mates,
they lay eggs. Hence the life cycle process begins again, so
that first phase comes again and the life cycle process
repeated. The adult stage is what most people think of when
they think of butterflies. The caterpillar and larva seems really
different from each other. The caterpillars have stubby
branches, very short antennae and little tiny eyes, etc. The
adult butterflies have long branches, long antennae, and
compound eyes. They can too fly by using their large and
colorful wings [16-17].
The behavior of adult butterflies during the food
search process is very realistic for finding the food plant. In
this process the butterflies find nectar sources or food plants
(flowers). The butterfly have different sensors such antenna,
eyes, etc., to find out the food plant. They can communicate or
exchange information between them and neighbors also by
various ways, such as dancing, color, chemicals, sound, and
physical actions, etc. By this butterfly shows the collective
intelligence behavior on the butterfly network.
The butterfly graphs were originally brought out as
the underlying graph of FFT networks, which can execute the
fast Fourier transform (FFT) very efficiently. The butterfly
wing structure is implemented by graph theory, then from
butterfly graph. The connection between two or more
butterflies graph form butterfly network. The butterfly
intelligence network (based on wing structure) is used to
represent the linear network also [25-27].

w: is the inertia weight.


rl , r2 : two random numbers between (0, 1).
Cl , C2 : the cognitive and social scaling parameters.
III.

THE BUTTERFLIES SWARM BEHAVIOR AND


INTELLIGENCE NETWORK

The butterflies as an advanced insects have a complete life


cycle. Butterflies pass through a life cycle. There are basically
four different phases. The first beginning phase is the eggs.
This is the phase in which eggs lays down by girl butterfly.
The butterfly places the eggs on leaves. After that, the second
phase start, its the caterpillar. This is the phase in which the
eggs incubate [21, -23]. It holds approximately five days for
the eggs to incubate, by completing this process, caterpillar
comes out. At this time, the caterpillar eats all the time. The
growth of the caterpillar is very fast. Once its grown all the
way, then the third stage starts. It is like the chrysalis stage of
the butterfly [12-15]. The caterpillar forms a chrysalis. The
caterpillar is inside the chrysalis. Within the chrysalis, it
begins to change. It fast changes into a butterfly. At one time
the caterpillar has changed into a butterfly, the fourth stage

Fig. 1. Butterfly network structure [25-27]

This butterfly network having different nodes A, B, C and D


as shown in figure 1. These vertex or node can be assumed as
communication nodes of butterfly network, which used to
exchange or transmit information between butterflies by
different means.
IV.

THE BUTTERFLY SWARM BASED SEARCH


PROCESS

In this section butterfly, intelligent network is formed when


the sensitivity of butterfly meets the probability of the nectar.
In the sensitive region with minimum sensitivity of butterfly
matches minimum probability of nectar. In this region, all

2014 First International Conference on Networks & Soft Computing

173

food sources are called in active state, while in the outer


region present inactive state. The outer region where
sensitivity of a butterfly is not able to detect the probability of
nectar in this region. The active and inactive regions are
separated by the boundary which is from when the minimum
sensitivity of butterfly meets to the minimum probability of
nectar. A network consists of N nodes having sourced of
nectar. Each node (food source) is selected by their degree,
when the node does not have nectar than the degree of a node
is considered zero, and when there is the nectar source than
according to sensing ability of butterflies connectivity will
increase the degree (higher degree) of a node in the sensitive
region because all the nodes outside the region will have a
lower degree. Depending of the degree we can classify or
arrange the nodes of the best nectar source. The optimal
solution in the sensitive region (probabilistic region) will
depend on the degree of the nodes. In the butterfly network
node (flower) having more amounts of nectar will attract more
and more butterflies by natural sensitivity; hence it will
increase the degree of the node (flower) because of that node
has a higher probability of nectar. The butterflies exchange the
information not only by dancing, but also they have natural
intelligence for sensing nectar through color, insect, chemical,
sound and physical action so it develops better communication
network compared to others. This intelligent communication
systems, that is the information exchange between the each
and every adult butterflies contributes and perform the
"collective intelligence" behavior within the butterfly network
[12-15]. The structure of the butterfly network consisting N
nodes at instant t, the butterfly communication network states
of regions is given in figure 2, which can be defined as:

given in figure 3. Dancing mechanism is not understood in


bees, but here assumption is that the recruitment amongst the
entire bee is the function of good-quality food, butterfly
develops a good communication network by several
parameters such as color, chemicals, sound, and physical
actions. They can it chooses the flower of maximum nectar
probability randomly search the optimal solution and before
terminating the respective flight (iteration) it finds a next
optimal solution. Hence the loop continues and also
propagates with the time, so in less time butterflies are able to
choose several local best (lbest) them according to
arrangement of degree of nodes, they choose selects global
best (gbest). During the forging process in new search region
decision making process butterfly interact with the
substantially different sensory system. So in case of butterfly
the search has somewhat become sequential, so it limits the
area, hence it takes less time compared to others. It searches
for nectar near the plant where it lays eggs. Butterflies don't
make nests like bees or wasps; usually they live on same food
source until it finds a new food source [15-19]. The algorithm
termination criteria will be the maximum number of flights
(iteration) with improvement in fitness of respective objective
function. Improvement of butterfly-PSO according to
ingeniously and due to antenna on their mouth by which
flowers attract butterfly more compare to others there is a very
good relation of butterfly and surrounding atmosphere to
provide the best solution. After butterfly relinquishes the food,
butterfly propagate in the direction of the new nectar source or
gets information from others.



 
(3)


 
(4)

Fig. 3. BF-PSO search process representation


Fig. 2. Active and inactive regions concept of BF-PSO

V.
In the food search process, butterfly finds the optimal location
depending upon the sensitivity of the flower and probability of
nectar, after finding the solution, it communicates directly or
indirectly from the others by different mean of communication
intelligence. The representation for BF-PSO search process is

174

THE BUTTERFLY PARTICLE SWARM BASED


OPTIMIZATION TECHNIQUES (BF-PSO)

The butterfly PSO algorithm is essentially based on nectar


probability factor and sound communication system means

2014 First International Conference on Networks & Soft Computing

sensitivity. BF-PSO consists of intelligent behavior of the


butterfly to find out the nectar in maximum quantity.
By taking the effect of sensitivity and nectar
probability we have modified the standard PSO with some
necessary variation in optimization parameters. The sensitivity
is more nearer to the nectar source, and the minimum
sensitivity is required to search the next nectar sources, so the
range of the sensitivity and probability from 0.1 to 1.

chosen. The standard benchmark functions as given in the


table I (the functions F1, F2, F3 and F4). The comparative
results for functions using various optimization methods such
as genetic algorithm (GA), S-PSO and BF-PSO given below
in figures.4, 5, 6 and figure 7. Also the optimum values for
each function with different dimensions are given in table II
and table II.
4

x 10

SPSO
BFPSO
GA

The parameters and equations for BF-PSO are given as:


The sensitivity is:
S = exp(-(itera_max-itera)/itera_max)

(5)

Fitness value

6
5
4
3
2
1
0

100

200

300

400

500
600
no. of iterations

700

800

900

1000

11

x 10

SPSO
BFPSO
GA

7
6
5
4
3
2
1

The probability of an event P(E) can be given as:

P(E) =

Fig. 4. Comparative results for function F1

Fitness Value

The Probability in an experiment or assignment for an event


evaluates how likely an event is. In other words, probability
measure the chance for that likely event occurs as a result of
an experiment [6]. The results of an experiment which leads to
chance or possibility of an event is called favorable outcomes.
The probability of an event is number of likely observations of
the events divide by the total number of observations in the
assignment. In any event, the probability that an event will
occur is lies between 0 and 1. The probability value 1 denotes
sure event, occurs every time in an experiment and it's
repeated. The probability value 0 denotes impossible event,
the event never occurs in an experiment.

100

total numberof favourableoutcomes fortheeventinanExperiment


total numberof possibleoutcomesinanExperiment

200

300

400
500
600
No. of iteration

700

800

900

1000
100

Fig. 5. Comparative results for function F2


90

SPSO
BFPSO
GA

80

(6)

70
60
Fitness value

In the similar way the probability of global best butterfly


swam is given as:

50
40
30

P=

global best fitness of BF swarm


N

i =1 local best fitness of all BF swarms

20

(7)

10
0

100

200

300

400

500
600
No. of iteration

700

800

900

1000

Fig. 6. Comparative results for function F3

Then modified equations are the basic equations of BF-PSO


technique given below for the velocity and position updating:

v '( k +1) = w * vk + sk (1 pk )c1 r1 (lbest k currentpop )

x ( k + 1 ) = x k + v 'k + 1
VI.

(8)
(9 )

SPSO
BFPSO
GA

2.5

2
Fitness value

+ pk c2 r2 ( gbestk currentpop )

x 10

1.5

RESULTS

The implemented results to validate the proposed BF-PSO


algorithm on the several well known functions have been

0.5

100

200

300

400

500
600
No. of itrration

700

800

900

1000

Fig. 7. Comparative results for function F4

2014 First International Conference on Networks & Soft Computing

175

TABLE I. THE BENCHMARK FUNCTION AND RANGES


Name
F1

Equation

Range

[2]

 

[-10, 10]

[3]

F3

F4

[-100, 100 ]

[-30, 30]

TABLE II. THE COMPARISON OF RESULTS FOR BENCHMARK


FUNCTIONS (1000 ITERATION AND 20 TRIALS)
GA

Algorithm

S-PSO

Mean
Best

Mean
Best

Mean
Best

30

7.50808e+
003

2.23034e006

1.85086e010

F2
F3

30
30

26.52993
40.96571

0.06236
0.84339

0.00337
0.21226

F4

30

2.08842e+
006

43.45286

25.2002

CONCLUSIONS

The effect of sensitivity of butterfly and probability of nectar


in BF-PSO proposed algorithm, for the various test functions
are improved results as compared to other methods (such as
GA, PSO etc). The search operation has been increased to a
larger extent. BF-PSO shows the good convergence rate and at
each time the algorithm provides more optimum output, hence
in BF-PSO results found with sound accuracy as well good
convergence.
In the existing work presented by Magnus Erik, and
Hvass Pedersen and Hvass Laboratories [24] for the
benchmark functions F1 (sphere), F2 (rosenbrock), F3
(schwefel2-22), F4(schwefel2-21). They have compared test
function results with PSO (Particle Swarm optimization) and
MOL (Many Optimizing Liaisons). The results given in the
table II gives the better results as compared to existing one. So
that due to introducing the two new parameters the sensitivity
of butterfly and probability of nectar in BF-PSO give the best
optimized values; hence there is a higher convergence rate in
the BF-PSO then other method. The convergence of fitness
towards minimum value with respect to iteration for BF-PSO
compared with S-PSO and GA is shown in figures from 4 to 7.
However the mean value of BF-PSO is somewhat higher
compared to PSO, so diversity increases, and consists of high
accuracy with higher convergence rate.

176

[5]

[6]

F1

VII.

[4]

BF-PSO

Dime
nsion
Function

[1]

[-100, 100 ]

F2

REFERRENCES

[7]

[8]

[9]

[10]

[11]
[12]

[13]

[14]

[15]

[16]

[17]

Wilson, E. O. Chemical communication within animal species In


Chemical ecology, Vol. 9, ed. E. Sondheimer & J. B. Simeone, pp.
133155. New York, Academic Press, (1970).
Tristram D. Wyatt, Pheromones and Animal Behaviour,
Communication by Smell and Taste Cambridge University of press,
2003.
J. Kennedy, and R.C. Eberhart, Particle Swarm Optimization, IEEE
International Conference on Neural Networks Proceedings, pp. 19421948, vol. 4, Dec. 1995.
Eberhart, R. C., and Kennedy, J. A new optimizer using particle swarm
theory. Proc. Sixth International Symposium on Micro Machine and
Human Science (Nagoya, Japan), IEEE Service Center, Piscataway, NJ,
39-43(1995).
J. Kennedy, The particle swarm: social adaptation of knowledge. Proc.
IEEE International Conference on Evolutionary Computation
(Indianapolis, Indiana), IEEE Service Center, Piscataway, NJ, pp. 303308, 1997.
Robert
B.
Ash,
Basic
Probability
Theory
DOVER,
PUBLICATIONS, INC., pp. 1-30, Mineola, New York, 2008.
Randy L. Haupt, Sue Ellen Haupt, Practical genetic algorithms
second edition, 2004, published by john wiley & sons, inc., hoboken,
new jersey, published simultaneously in Canada, 2004.
M. Falahi Sohi & M. Shirdel, "Applying BCO Algorithm to Solve the
Optimal DG Placement and Sizing Problem," Elect. Electron. Eng., vol.
2, pp. 3137, 2012.
T. Tsujimoto, T. Shindo & T.Kimura, "A Relationship between
Network Topology and Search Performance of PSO," WCCI 2012,
IEEE World Congress on Computational Intelligence, June, 10-15,
2012 - Brisbane, Australia.
B.K. Babayigit, R. Ozdemir &, "A modified artificial bee colony
algorithm for numerical function optimization," IEEE, 2012. IEEE
Symposium on Computers and Communications (ISCC), 2012 , pp.
000245 000249.
J. Kennedy, Bare Bones Particle Swarms, Proceedings of the IEEE
Swarm Intelligence Symposium, 2003, SIS-03, pp. 8087, 2003.
Jacques Demongeot, Michel Morvan, Sylvain Sene, Impact of Fixed
Boundary Conditions on the Basins of Attraction in the Flowers
Morphogenesis of Arabidopsis Thaliana IEEE, 22nd International
Conference on Advanced Information Networking and Applications,
Workshops, WAINA, 2008, pp. 782-789.
Debbie Leung, Jonathan Oppenheim, and Andreas Winter Quantum
Network Communication-The Butterfly and Beyond IEEE
Transactions on Information Theory, vol. 56, no. 7, pp. 3478-3490,
july 2010.
Shigeru Sunada, Keiji Kawachi, Isao Watanabe And Akira Azuma,
Performance of a butterfly in take-off flight The J. Experimental
Biology, 183, pp. 249277, June 1993.
R. J. Mercader, L. L. Stelinski and J. M. Scriber, Differential antennal
sensitivities of the generalist butterflies papilio glaucus and p.
canadensis to host plant and non-host plant extracts Journal of the
Lepidopterists Society, 62(2), pp. 84-88, 2008.
Kathleen M. Lucas, James F. C. Windmill, Daniel Robert and Jayne E.
Yack, Auditory mechanics and sensitivity in the tropical butterfly
Morpho peleides (Papilionoidea, Nymphalidae) The J. Experimental
Biology, 212, pp. 3533-3541, July 2009.
Douglas Blackiston, Adriana D. Briscoe and Martha R. Weiss, Color
vision and learning in the monarch butterfly, Danaus plexippus
(Nymphalidae) The J. Experimental Biology, 214, pp. 509-520,
October 2010-11.

2014 First International Conference on Networks & Soft Computing

[18] John Paul Cunningham1, Chris J. Moore, Myron P. Zalucki and Stuart
A. West Learning, odour preference and flower foraging in moths
The J. Experimental Biology, 207, pp. 87-94, September 2003.
[19] Kathleen M. Lucas, James F. C. Windmill, Daniel Robert and Jayne E.
Yack, Auditory mechanics and sensitivity in the tropical butterfly
Morpho peleides (Papilionoidea, Nymphalidae) The J. of
Experimental Biology, 212, 3533-3541, July 2009.
[20] Norouzzadeh, M.S.; Ahmadzadeh, M.R.; Palhang, M., "Plowing PSO:
A novel approach to effectively initializing particle swarm
optimization," Computer Science and Information Technology
(ICCSIT), 2010 3rd IEEE International Conference on , vol.1, no.,
pp.705,709, 9-11 July 2010.
[21] Lucic, P.; Teodorovic, D., "Transportation modeling: an artificial life
approach," Tools with Artificial Intelligence, 2002. (ICTAI 2002).
Proceedings. 14th IEEE International Conference on , vol., no.,
pp.216,223, 2002.
[22] Dusan Teodorovic, Swarm intelligence systems for transportation
engineering: Principles and applications Transportation Research Part
C: Emerging Technologies, Volume 16, Issue 6, pp. 651667,
December 2008.
[23] Yue-Jiao Gong; Meie Shen; Jun Zhang; Kaynak, O.; Wei-Neng Chen;
Zhi-Hui Zhan, "Optimizing RFID Network Planning by Using a
Particle Swarm Optimization Algorithm With Redundant Reader
Elimination," Industrial Informatics, IEEE Transactions on , vol.8,
no.4, pp.900,912, Nov. 2012.
[24] Magnus Erik, Hvass Pedersen, Hvass Laboratories, Good Parameters
forParticle Swarm Optimization Technical Report, no. HL1001, pp. 1-

12, 2010. doi. http://www.cof.orst.edu/cof/teach/fe640/Class_Materials/


Particle%20Swarm/PSO%20parameters.pdf
[25] http://www.canisius.edu/~grandem/butterflylifecycle/
The_Lifecycle_of_a_Butterfly_print.html
[26] http://www.thebutterflysite.com/life-cycle.shtml
[27] http://en.wikipedia.org/wiki/Butterfly

Aashish Kumar Bohre1 received BE degree in Electrical and Electronics


engineering from UIT- RGPV Bhopal, (2009), and M-Tech degree in Power
System (2011) from MANIT, Bhopal. At the moment he is Pd.D. scholar at
MANIT,
Bhopal,
India.
Email:
aashu371984@gmail.com,
aashish.bohre@gmail.com
Dr. Ganga Agnihotri2 received BE degree in Electrical engineering from
MACT, Bhopal (1972), the ME degree (1974) and PhD degree (1989) from
University of Roorkee, India. Since 1976 she is with Maulana Azad College
of Technology, Bhopal in various positions. Currently she is professor. Her
research interest includes Power System Analysis, Power System
Optimization and Distribution Operation. Email: ganga1949@gmail.com
Dr. Manisha Dubey3 was born in Jabalpur in India on 15th December 1968.
She received her B.E (Electrical), M.Tech. (Power Systems) and Ph.D
(Electrical Engg.) in 1990, 1997 and 2006 respectively. She is working as
Professor at the Department of Electrical Engineering, National Institute of
Technology, Bhopal, India. Her research interests include power systems,
Genetic Algorithms, Fuzzy Logic systems and application of Soft Computing
Techniques in power system dynamics and control. Email:
manishadubey6@gmail.com

TABLE III. THE RESULTS FOR BENCHMARK FUNCTIONS (100000 ITERATION AND 100 TRIALS)
Algorithm

Dimensi
on

Best

S-PSO
Mean

Std. dev.

Best

BF-PSO
Mean

Std. dev.

2
5
10
20

0
0
2.6322e-243
1.4948e-107

0
0
7.8328e-228
2.5578e-093

0
0
0
2.5410e-092

0
0
0
3.6856e-149

0
0
0
8.4820e-095

0
0
0
8.4820e-094

30
2
5
10
20
30
2
5
10
20
30
2
5
10
20
30

6.9550e-060
0
9.6438e-241
2.5704e-136
1.6966e-062
4.2881e-037
0
7.4581e-187
1.0425e-059
8.5961e-010
0.0046
0
3.7325e-008
8.3947e-008
0.0145
0.0711

2.2670e-057
0
9.1180e-230
5.3000
6.2000
18
0
1.5974e-175
1.2085e-052
6.0935e-007
0.0883
0.2004
1.2020
3.8031e+003
1.9121e+003
3.7969e+003

4.8193e-050
0

1.3530e-066
0
0
5.7095e-242
5.2522e-197
1.9579e-041
0
0
3.2675e-135
7.7867e-024
1.2286e-007
0
5.4141e-012
4.9854e-011
1.4295e-008
5.6642e-004

4.5480e-058
0
0
2.4000
1
18
0
3.0287e-317
1.3719e-118
2.5853e-016
6.0233e-005
0
0.7439
1.6312e+003
9.6327e+002
3.6306e+003

1.3459e-052
0
0
5.8810
3.1623
25.7337
0
0
1.3719e-117
1.6592e-015
1.5112e-004
0
3.8281
1.7722e+004
9.00e+003
1.7721e+004

Function

F1

F2

F3

F4

8.7145
13.9102
26.9680
0
0
5.8852e-052
3.8596e-006
0.1935
2.0044
4.8593
3.2663e+004
1.2659e+004
1.7968e+004

2014 First International Conference on Networks & Soft Computing

177

You might also like