See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/232608310

Improved Local Search in Artificial Bee Colony using Golden Section Search

Article · October 2012

Source: arXiv

CITATIONS READS
41 270

3 authors, including:

Dr. Tarun Kumar Sharma Millie Pant

Shobhit Institute of Engineering and Technology Indian Institute of Technology Roorkee
133 PUBLICATIONS 1,311 CITATIONS 278 PUBLICATIONS 6,229 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Dr. Tarun Kumar Sharma on 15 October 2014.

The user has requested enhancement of the downloaded file.

 Improved Local Search in Artificial Bee Colony using Golden
Section Search
1
Tarun Kumar Sharma, 2Millie Pant, 3V.P.Singh
1, 2
Indian Institute of Technology Roorkee, Roorkee, India
3
SCET, Saharanpur, India
Email:{taruniitr1; millidma, singhvp3}@gmail.com

Abstract –Artificial bee colony (ABC), an optimization algorithm is a recent addition to the family of population based
search algorithm. ABC has taken its inspiration from the collective intelligent foraging behavior of honey bees. In this
study we have incorporated golden section search mechanism in the structure of basic ABC to improve the global
convergence and prevent to stick on a local solution. The proposed variant is termed as ILS-ABC. Comparative
numerical results with the state-of-art algorithms show the performance of the proposal when applied to the set of
unconstrained engineering design problems. The simulated results show that the proposed variant can be successfully
applied to solve real life problems.
Keywords –Artificial Bee Colony; ABC; Optimization; Golden Section; Metaheuristics;

1. Introduction proposed to achieve these two goals [21]-[27]. The

comprehensive survey of ABC can be found in [28]. In
Population based search heuristics like Genetic the present study we improved the movement of onlooker
Algorithm (GA) [1], Particle Swarm Optimization (PSO) bees using local search method called golden section
[2], Differential Evolution (DE) [3], Ant Colony search method in order to balance exploration &
Optimization (ACO) [4] and a like have attracted the exploitation and to get more efficient food locations. The
researchers and scientist to solve complex problems proposed ILS-ABC is tested four engineering design
arising in the domain of engineering design, finance, problems which show its performance over basic ABC.
chemical etc. Population based search techniques are The rest of the paper is organized as follows: In
nontraditional search techniques which do not require any section 3, the proposed variant ILS-ABC is described.
auxiliary properties like differentiability and continuity of The problem set taken to validate the proposal is
the objective function as well as also independent of the mentioned in section 4. Experimental settings,
nature of the problems. These algorithms can be comparison criteria and result analyses are given in
successfully applied to solve many real life problems [5]- section 5. The conclusions are presented in section 6.
[7].
Artificial Bee Colony (ABC), proposed by Karaboga 2. Artificial Bee Colony: A Brief Overview
[8]-[10] is the recent addition to the population based
search heuristics that is inspired by the intelligent Division of labor and self organization are the
foraging behavior of honey bees. ABC comprises of three component keys in bee colony. In a self-organizing
kinds of bees namely (a) Scout (b) Employed and (c) system, each of the covered units may respond to local
Onlooker bees. The bees intelligently organize stimuli individually and act together to accomplish a
themselves and divide the labor to perform the tasks, like global task via division of labor without any centralized
searching for the nectar, sharing the information about the regulation. According to [29] positive feedback, negative
food source etc. The position of a food source represents feedback, multiple interactions and fluctuations are the
a possible solution to the optimization problem and the four characteristics on which self-organization rely.
nectar amount of a food source corresponds to the quality Foraging, nest building, marriage, task selection and
(fitness) of the associated solution. An epigrammatic navigation are the few tasks that bee swarms performs.
mathematical narration of ABC is presented in section 3. Task selection is dependent on the environment and hive,
Like other population based search heuristics ABC has which can be changed adaptively. Foraging is one of the
gained a great attention of researchers and scientists to major tasks for the bees. There are three types of bees
implement it to solve many real life problems in versatile associated with the foraging task, employed, onlooker
domain. The latest applications of ABC can be found in and scout bees. Similarly ABC tries to model natural
[11]-[20]. A brief overview of ABC is given in section 2. behavior of real honey bees in food foraging. The colony
Like other evolutionary algorithms ABC also has of artificial bees also contains three groups of bees:
some drawbacks which obstruct its performance. ABC is employed bees which are responsible for exploiting the
good at exploration while poor at exploitation. Therefore, food sources and pass the information to the onlooker
accelerating convergence speed and avoiding the local bees, which are waiting in the hive. Onlooker bees
optima have become two important and appealing goals chooses the food sources by watching the waggle dance
in ABC research. A number of ABC variants have, been performed by employed bees while scouts explores the
 2

food sources randomly based on some external clue [30]. randomly to SN number of employed bees and their
This intelligent foraging mechanism can be better fitness is determined.
explained using the graphical representation shown in
Employed Bees Stage
figure 1.
The duty of employed bees is to determine new food
source say, vi with the help of the food source xi assigned
to it during the initialization phase. The equation used is:
vij  x ij  ij ( xij  x kj ) (2)

where k  {1,2,…,SN} and j  {1,2,…,n}; Control

parameter ij  [-1,1] gives the step length which decides
the movement of bees. k is generated randomly however,
k ≠ i. After a new food source has been generated by the
employed bee, a tournament selection is held between vi
Waggle dance and xi and the one with a better fitness value becomes the
member of the population.
Figure 1. Graphical representation of the elements of ABC algorithm
Onlookers Phase
In the initial phase of foraging, bee explores the The work of onlookers is to enhance the exploration
environment randomly in search of food sources. When capability of ABC. An onlooker evaluates the nectar
the forager bee finds the food source it becomes information (i.e. the fitness of solutions) collected from
employed bee. After exploiting the discovered food all the employed bees and selects a food source xi using
source, employed bee returns to the hive and unloads the probability Pi as:
gathered nector. Now it’s up to bee to go back directly to
SN
the discovered food source or can share the site Pi  f i /  f k (3)
k 1
information to the onlooker bees waiting in the hive by
performing a dance called waggle dance on the dance where fi denotes the fitness value of the ith food source.
area. Through this dance the employed bee informs the The onlooker after selecting the food source xi, modifies
onlooker bee about the direction w.r.t. Sun, distance and it by using Equation (2). Once again a tournament is held
quality (Fitness) of the food source. If the food source between the food source selected and the food source
exploited by employed bee gets exhausted then the bee modified by the onlooker and once again as in previous
becomes scout bee, also called the replacement step, the one with better fitness value becomes a member
mechanism, and explores the new food sources randomly. of the population.
The Onlooker bees choose the source site depending on
the frequency of the dance, as frequency is proportional Significance of Scouts
to quality of the food source. This is how the onlooker The main work of scouts is to induce diversity in ABC.
bee becomes the employed bee. An employed bee which is not able to improve the fitness
In ABC algorithm the food Sources represents the value of a food source becomes a scout. A scout is
candidate solutions/possible solutions and the nector activated when it is observed that the fitness of a food
amount of the food source represents the fitness source xi shows no betterment even after a specified
associated with the solution for the optimization problem. number of trials limit. Stagnation in the fitness value of a
And each of the food sources is exploited by only one food source indicates that the particular food source may
employed bee hence the numbers of food sources are be replaced with a new food source. Scout produces a
equal to the employed bees. new food source with the help of equation (1).
By this analogy between the intelligent foraging
behavior of bees and the ABC algorithm, the basic ABC Pseudocode of the ABC:
algorithm can be explained as follows: Initialization of the Food Sources
Initialization: Evaluation of the Food Sources
Repeat
Like all other population based search heuristics, the Produce new Food Sources for the employed bees
first step of ABC is generation of initial population. We Apply the greedy selection process
assume that the initial population is made up of SN Calculate the probability values for Onlookers
number of n-dimensional real-valued vectors, Produce the new Food Sources for the onlookers
representing the food sources, which are generated Apply the greedy selection process
randomly. Suppose X i  {xi ,1 , xi , 2 ,...,xi ,n } denotes the ith Send randomly scout bees
food source in the population, where each food source is Memorize the best solution achieved so far.
initialized as. Until termination criteria are met.

xi, j  xmin, j  rand(0,1)(xmax, j  xmin, j ) (1)

here i  1,2,..,SN; j  1,2,…,n. xmax,j and xmin,j denotes the 3. Local Search in ABC: ILS-ABC
upper and lower bound constraints respectively of
decision variable. These food sources are assigned
 3

In order to improve the performance of basic ABC, The performance of ILS-ABC is evaluated on test bed
golden section search [31] is introduced in the structure of four engineering design real life problems taken from
of basic ABC. This process helps in balancing the literature.
exploration and exploitation. The poor balance may result The performance of proposed algorithms is evaluated
in weak optimization method which may suffer from on a test bed of seven unconstrained real life problems
trapping in a local optima, stagnation and premature that are common in various fields of engineering designs.
convergence. Introduction of golden section search helps These are: (A) Optimization of Transistor Modeling (B)
in improving the global convergence and to prevent to Optimal capacity of gas production facility (C) optimal
stick on a local food source of the ABC. Hence the thermohydraulic performance of an artificially roughened
proposed variant is termed as ILS-ABC. In the proposed air heater and (D) Design of a gear train. All the problems
algorithm the onlooker bee replaces the food sources in considered in the present study are highly nonlinear in
basic ABC (Equation 2) using the following Equation: nature. Mathematical models of the problems are
vij  x ij  Fz ( xij  x kj ) (4) discussed below:
where Fz is scaling factor depends on the value of a & b (A) Optimization of transistor modeling [32]: The
discussed below. The generation of a new food sources is objective function of this problem provides a least-
seen as an opaque procedure depending on the scale sum-of-squares approach to the solution of a set of
factor Fz. nine simultaneous nonlinear equations, which arise
Scaling factor becomes an important aspect that is to in the context of transistor modeling. The
be controlled in order to guarantee a high quality food mathematical model of the transistor design is given
source (solution) that can have an important role in the by,
4 2 2
succeeding generations. Minimize f ( x)   2   ( k   k )
k 1
A local search procedure is applied to Fz in order to
Where
detect the scale factor value which guarantees a food
 k  (1  x1 x 2 ) x 3 {exp[ x5 ( g1k  g 3k x7  10 3  g 5k x8  10 3 )]  1}g 5 k  g 4 k x 2
sources with a high performance. This can be gained by
 k  (1  x1 x 2 ) x4 {exp[ x6 ( g 1k  g 2 k  g 3 k x 7  10 3  g 4 k x9  10 3 )]  1}g 5 k x1  g 4 k
minimizing the function f ( Fz ) in the decision space
  x1x3  x2 x4
[−1, 1]. The meaning of the negative scale factor is the
inversion of the search direction. If the search in the Subject to: xi  0 , i = 1, 2, … , 9
negative direction succeeds, the corresponding positive And the numerical constants g ik are given by the matrix
value is assigned to the food sources for the subsequent
 0.485 0.752 0.869 0.982 
generations.  
The Scale factor golden section search applies the  0 .369 1 . 254 0 .703 1.455 
Golden section search to the scale factor in order to  5.2095 10.0677 22.9274 20.2153 
generate a high quality food sources. This scheme  
 23.3037 101.779 111.461 191.267 
processes in the interval [a = −1, b = 1] and generates two  28.5132 111.8467 134.3884 211.4823
intermediate points:  
ba ba
Fz1  b  ; Fz2  a  (B) Optimal capacity of gas production facilities [33]:
  This is the problem of determining the optimum
1 5 capacity of production facilities that combine to
where   is the golden ratio. The evaluated make an oxygen producing and storing system. The
2 mathematical model of this problem is given by,
values of f ( Fz1 ) and f ( Fz2 ) are then compared and if x2  0.85
f ( x)  61.8  5.72 x1  0.2623[(40  x1 ) ln( )] 
f ( Fz1 ) < f ( Fz2 ) then Fz2 replaces b and this procedure is Minimize 200
repeated in the new smaller interval [a, b]. The process is x2  0.75
0.087(40  x1 ) ln( )  700.23x2
described as below: 200

Repeat Subject to: x1  17.5 , x2  200

ba ba
Compute Fz1  b  ; Fz2  a 
  Bounds: 17.5  x1  40 , 300  x 2  600 .
Evaluate f ( Fz1 ) and f ( Fz2 )
(C) Optimal thermohydraulic performance of an
If f ( Fz1 ) < f ( Fz2 ) then artificially roughened air heater [34]: In this
b = Fz2 problem the optimal thermohydraulic performance of
an artificially roughened solar air heater is
Else
considered. Optimization of the roughness and flow
a = Fz1 parameters (p/e, e/D, Re) is considered to maximize
End If the heat transfer while keeping the friction losses to
Until termination criteria are met. be minimum. The mathematical model of this
problem is given by,
4. Test Bed: Real life problems
Maximize f ( x )  2.51* ln e   5.5  0.1RM  GH
 4

Where R M
 0.95 x 2
0.53
; GH  4.5(e ) 0.28 (0.7) 0.57 , implies the mean of the best fitness values obtained by

e  x1x3 ( f / 2) 1/ 2
f  ( f s  f r ) / 2 f s  0.079x3 0.25 the algorithm till the stopping criteria is satisfied,
; ; ; Number of function evaluations (NFE) tells about the
0.53
f r  2(0.95 x3  2.5 * ln(1 / 2 x1 ) 2  3.75) 2 speed of an algorithm. Smaller NFE shows that the
;
algorithm converges more quickly. Numerical results
Subject to: 0.02  x1  0.8 , 10  x 2  40 , 3000  x3  20000 based on these performance measures are given in Table
(D) Design of gear train [32]: This problem is to 1 and Table 2. From Table 1, which gives the average
optimize the gear ratio for the compound gear train. fitness function value, NFE and execution time, we see
It is to be designed such that the gear ratio is as close that in terms of average fitness function value and time
as possible to 1/6.931. For each gear the number of all the algorithms gave more or less similar results
teeth must be between 12 and 60. Since the number although in some cases the proposed algorithms gave a
of teeth is to be an integer, the variables must be marginally better performance than PSO, DE and ABC.
integers. The mathematical model of gear train However if we compare the NFE in Table 2 the superior
design is given by, performance of the proposed algorithms become more
evident.
2 2
 1 T T   1 xx  Acceleration rate (AR) [35] is used to compare the
Minimize f ( x)    d b   1 2 convergence speeds between ILS-ABC and other
 6. 931 T T
a f   6. 931 x x
3 4 
algorithms. It is defined as follows:
Subject to: 12  xi  60 i  1, 2,3, 4 NFE one a lg orithm  NFE other a lg orithm
AR  %
[ x1 , x 2 , x3 , x 4 ]  [Td , Tb , Ta ,T f ] , xi’s should be integers. Ta, NFE one a lg orithm
Tb , Td, and Tf are the number of teeth on gears A, B, D From Table 2 we can see that the proposed ILS-ABC
and F respectively. gives the better results for every problem in the
comparison to the PSO, DE and ABC. Further from the
Table 2 it is clearly analysed that the proposed ILS-ABC
is faster than PSO by 38.6%, DE by 27.7% and faster
than ABC by 31.6%.
Table 1. Simulation results of four engineering design real life
problems
Figure 2. Design of Gear Train
Source
Item PSO DE ABC ILS-ABC
Result
5. Experimental settings, Result analyses and (A) Optimization of Transistor Modeling
x1 0.9010 0.9013 0.9009 0.9002 0.90
Comparison criterion x2 0.8841 0.5878 0.5224 0.4519 0.45
x3 4.0386 3.7812 1.0764 1.0352 1.0
5.1. Experimental Settings x4 4.1488 3.9021 1.9494 2.045 2.0
x5 5.2436 5.1962 7.8536 5.495 8.0
x6 9.9326 11.2697 8.8364 9.734 8.0
ABC and the proposed variant ILS-ABC are x7 0.1009 0.0979 4.7712 0.103 5.0
implemented on Dev-C++ and the experiments are x8 1.0599 1.1053 1.0074 1.012 1.0
conducted on a computer with 2.00 GHz Intel (R) core x9 0.8066 0.6799 1.8545 0.610 2.0
(TM) 2 duo CPU and 2- GB of RAM. For each problem, f(x) 0.0695 0.0618 0.0113 0.0304 NA
NFE 22195 19784 17901 14932 NA
all the algorithms independently run 25 times. Time 0.863 0.415 0.401 0.345 NA
The population size, MCN (maximum cycle numbers) (B) Optimal Capacity of Gas Production Facilities
and limit are fixed to 40, 10000 and 100 respectively. x1 17.5 17.5 17.5 17.5 17.5
And the reported results are the means and standard x2 600 593 600 600 465
deviations of the statistical experimental data. The f(x) 169.844 169.996 169.012 169.731 173.76
NFE 342 324 319 264 NA
random numbers are generated using inbuilt rand () Time 0.02 0.01 0.02 0.01 NA
function with same seed for every algorithm. (C) Optimal Thermo hydraulic Performance of an Artificially
For each algorithm, the stopping criteria is to Roughened Air Heater
terminate the search process when one of the following x1 0.05809 0.18005 0.029589 0.07111 0.052
conditions is satisfied: (i) the maximum number of x2 10 10 9.012 9.876 10
x3 10400.2 4918 5028.4 4220 10258
generations is reached (assumed 1000 generations), f(x) 4.2142 4.0923 4.2166 4.0703 4.182
(ii) | fmax - fmin | < 10-4 where f(x) is the value of objective NFE 6207 5190 4425 3026 NA
function. Time 0.3 0.3 0.3 0.2 NA
(D) Design of Gear Train
5.2. Result Analyses and Comparison Criteria x1 13 16 19 18.75 18
x2 31 19 16 13.6 22
x3 57 49 44 44 45
The performances of proposed ILS-ABC is x4 49 51 49 52 60
demonstrated on a set of four nonlinear engineering f(x) 9.98e-11 2.78e-11 2.78e-11 5.59e-07 5.7e-06
Gear ratio 0.14429 0.14428 0.14428 0.14424 0.14666
design problems and the numerical results are compared
Error (%) 0.007398 0.000497 0.000447 0.000401 1.65
with the basic PSO, DE, and ABC. Standard performance NFE 480 340 648 252 NA
measures like average fitness function values, average Time 0.1 0.01 0.02 0.01 NA
number of function evaluations and time taken to
simulate are calculated. Average fitness function value
 5

[5] Y. Zhang, L. Wu, Artificial Bee Colony for Two Dimensional

Protein Folding, Advances in Electrical Engineering Systems,
25000
1:1(2012) 19-23.
PSO DE ABC ILS-ABC
[6] Z. Zhang, Pattern Recognition by PSOSQP and Rule based
20000
System, Advances in Electrical Engineering Systems, 1:1(2012)
NFE
15000 30-34.
[7] H. Noormohamadi, A.A. Gharaveisi, M. Suresrafil, Novel
10000
Bacterial Foraging Algorithm for Optimization Problems,
5000 Advances in Electrical Engineering Systems, 1:1(2012) 41-48.
0
[8] D. Karaboga, An idea based on honey bee swarm for numerical
A B C D optimization. Technical Report-TR06, Kayseri, Turkey: Erciyes
Problems University; 2005.
[9] D. Karaboga, B. Basturk, A powerful and efficient algorithm for
(a) numerical function optimization: artificial bee colony (ABC)
algorithm, Journal of Global Optimization 39(2007) 171–459.
0.9
PSO DE ABC ILS-ABC [10] D. Karaboga, B. Basturk, On the performance of artificial bee
0.8
colony (ABC) algorithm, Applied Soft Computing, 8(2008) 687-
0.7
0.6
697.
0.5 [11] D. Karaboga, C. Ozturk, N. Karaboga, B. Gorkemli, Artificial bee
NFE

0.4 colony programming for symbolic regression, Information

0.3
Sciences (2012), http:// dx.doi.org/10.1016/j.ins.2012.05.002.
0.2
0.1 [12] M Ma, J. Liang, M. Guo, Y. Fan, Y. Yin, SAR image
0 segmentation based on Artificial Bee Colony algorithm, Applied
A B C D
Soft Computing, (In Press) doi:10.1016/j.asoc.2011.05.039, in
Problems
press.
(b) [13] WC. Yeh, TJ. Hsieh, Artificial bee colony algorithm-neural
networks for s-system models of biochemical networks
Figure 1. (a) NFE and (b) time taken to compile the four (A, B, C and
approximation. Neural Comput Appl. (2012) doi:10.1007/s00521-
D) engineering design real life problems. 010-0435-z.
[14] F. Gao, Feng-xia Fei, Q. Xu, Y. fang Deng, Yi-bo Qi, I.
Table 2. Comparison of ILS-ABC with PSO, DE and ABC in terms of Balasingham, A novel artificial bee colony algorithm with space
Acceleration rate (%) contraction for unknown parameters identification and time-delays
PSO Vs DE Vs ABC Vs of chaotic systems, Appl. Math. Comput. (2012),
Problems
ILS-ABC ILS-ABC ILS-ABC http://dx.doi.org/10.1016/j.amc.2012.06.040.
A 32.7 24.5 16.6 [15] H. Zhang, Y. Zhu, W. Zou, X. Yan, A hybrid multi-objective
B 22.8 18.5 17.2 artificial bee colony algorithm for burdening optimization of
C 51.2 41.7 31.6 copper strip production, Applied Mathematical Modelling,
D 47.5 25.9 61.1 36:6(2012) 2578-2591.
Average AR (%) 38.6 27.7 31.6 [16] X. Liao, J. Zhou, R. Zhang, Y. Zhang, An adaptive artificial bee
colony algorithm for long-term economic dispatch in cascaded
hydropower systems, International Journal of Electrical Power &
6. Conclusions Energy Systems, 43:1(2012) 1340-1345.
[17] H. Gozde, M. Cengiz Taplamacioglu, İ. Kocaarslan, Comparative
In this paper we present an algorithm ILS-ABC based performance analysis of Artificial Bee Colony algorithm in
on local search mechanism. In ILS-ABC golden section automatic generation control for interconnected reheat thermal
method is introduced in the onlooker phase of basic ABC, power system, International Journal of Electrical Power & Energy
Systems 42:1(2012) 167-178.
to improve the local search ability as well as tried to
[18] Ali R. Yildiz, A new hybrid artificial bee colony algorithm for
balance the two common and antagonist aspects, robust optimal design and manufacturing, Applied Soft
exploration and exploitation. The proposed algorithm is Computing, In Press, 2012.
applied to four real life engineering design problems. And [19] O. Kisi, C. Ozkan, B. Akay, Modeling discharge–sediment
the simulated results in terms of acceleration rate that relationship using neural networks with artificial bee colony
algorithm, Journal of Hydrology, 428–429(2012) 94-103.
directly indicates the better convergence speed, prove the
[20] S. K. Mandal, Felix T.S. Chan, M.K. Tiwari, Leak detection of
efficiency and applicability of the proposed work. pipeline: An integrated approach of rough set theory and artificial
bee colony trained SVM, Expert Systems with Applications,
39:3(2012) 3071-3080.
References [21] G. Li, P. Niu, X. Xiao, Development and investigation of efficient
artificial bee colony algorithm for numerical function
[1] D. Goldberg, Genetic Algorithms in Search Optimization and optimization. Applied Soft Computing 12(2012) 320– 332.
Machine Learning, Addison Wesley Publishing Company, [22] A. Bahriye, D. Karaboga, A modified Artificial Bee Colony
Reading, Massachutes 1986. algorithm for real-parameter optimization. Information Sciences
[2] J. Kennedy and R. C. Eberhart, Particle Swarm Optimization, 192(2012) 120-142.
Proceeding of IEEE International Conference on Neural [23] WF. Gao, S. Liu, L. Huang, A global best artificial bee colony
Networks, Perth, Australia, IEEE Service Center, Piscataway, NJ algorithm for global optimization. Journal of Computational and
(1995) pp. 1942–1948. Applied Mathematics 236(2012) 2741-2753.
[3] K. Price and R. Storn, Differential Evolution – a Simple and [24] T.K. Sharma, M. Pant, Enhancing the food locations in an
Efficient Adaptive Scheme for Global Optimization Over artificial bee colony algorithm, in: IEEE Swarm Intelligence
Continuous Spaces, Technical Report, International Computer Symposium (SIS), Paris, France, 2011, pp. 119-123.
Science Institute, Berkley 1995.
[25] T.K. Sharma, M. Pant, Enhancing different phases of artificial bee
[4] M. Dorigo, V. Maniezzo, A. Colorni, Positive feedback as a colony for continuous global optimization problems, in:
search strategy, Technical Report 91-016, Politecnico di Milano, International Conference on Soft Computing for Problem Solving,
Italy, 1991. SocProS 2011, AISC of Advances in Intelligent and Soft
Computing, Roorkee, India, Vol. 130, 2011, pp. 715–724.
 6

[26] T.K. Sharma, M. Pant, J.C. Bansal, Artificial Bee Colony with Conference on Bio-Inspired Computing: Theories
Mean Mutation Operator for Better Exploitation, in: IEEE World
Congress on Computational Intelligence (CEC), Brisbane,
and Application (BIC-TA 2012) and SoCProS-2012
Australia, 2012, pp. 3050 - 3056. (International Conference on Soft Computing for
[27] T.K. Sharma, M. Pant, J.C. Bansal, Some Modifications to Problem Solving).
Enhance the Performance of Artificial Bee Colony, in: IEEE
World Congress on Computational Intelligence (CEC), Brisbane, V.P. Singh is Director, SCET Saharanpur, and retired
Australia, 2012, pp. 3454 - 3461.
Professor from the department of Paper Technology, IIT
[28] D. Karaboga, B. Gorkemli, C. Ozturk, N. Karaboga: A
comprehensive survey: artificial bee colony (ABC) algorithm and
Roorkee. His research area includes Mathematical
applications, Artif Intell Rev 2011, DOI 10.1007/s10462-012- Simulation and Modeling. He has number of publications
9328-0. in journal of repute.
[29] E. Bonabeau, M. Dorigo G. Theraulaz, Swarm Intelligence: From
Natural to Artificial Systems, Oxford University Press, Inc., New
York, NY, USA, 1999.
[30] T.D. Seeley, The Wisdom of the Hive: The Social Physiology of
Honey Bee Colonies, Harvard University Press, 1995.
[31] J. Kiefer, “Sequential minimax search for a maximum, in
Proceedings of the American Mathematical Society, 4, 1953, pp.
502–506.
[32] B. Babu, New optimization techniques in engineering. Springer-
Verlag, Berlin Heidelberg, 2004.
[33] C. Beightler, D. Phillips. Applied geometric programming. John
Wiley and Sons, New York, 1976.
[34] B. Prasad, J. Saini. Optimal thermo hydraulic performance of
artificially roughened solar air heater, Journal Solar Energy,
47(1991) 91–96.
[35] S. Rahnamayan, H.R. Tizhoosh, and Salama, M. M. Ali,
Opposition-Based Differential Evolution. IEEE Transactions on
Evolutionary Computation, 12:1(2008) 64– 79.

Tarun Kumar Sharma did his

MCA in 2001, M.Tech (IT) in 2009
and presently pursuing Ph.D from
Indian Institute of Technology (IIT)
Roorkee, India. He has almost 9
years of teaching experience in
Engineering College. His key areas
are Evolutionary Computing;
Software Engineering; Computer
based Optimization Techniques; ERP. His research
interest includes swarm intelligence algorithms and their
applications in various complex engineering design
problems. His publications are in Journals and
International Conferences of repute. He volunteered in
SocPros-2011, an First International Conference on Soft
Computing for Problem solving. He is peer reviewer of
many IEEE conferences and International Journals. He is
student member of Machine Intelligence Research (MIR)
Labs, WA, USA.

Millie Pant is working as an

assistant professor in Department of
Paper Technology, Indian Institute
of Technology (IIT), Roorkee, India
since 2007. Her research interest
includes evolutionary and swarm
intelligence algorithms and their
applications in various complex engineering design
problems. Her publications are in Journals and
International Conferences of repute. She has published
over 100 referred on evolutionary algorithms (GA, PSO,
DE and ABC) and their applications in electrical design
problems, image processing papers. She has been
program committee member of over 10 International
events and Program Committee Chair of SoCProS-211.
She is Program Committee Chair of the 7th International

View publication stats