Proceedings of the International MultiConference of Engineers and Computer Scientists 2014 Vol I,

IMECS 2014, March 12 - 14, 2014, Hong Kong

Virtual Machine Scheduling Management on

Cloud Computing Using Artificial Bee Colony
B. Kruekaew and W. Kimpan

take advantages from these technologies without the

Abstract—Resource scheduling management design on requirements of deep knowledge or expertise. The Cloud
Cloud computing is an important problem. Scheduling model, aims to reduce costs and to provide the ease of resource
cost, quality of service, time, and conditions of the request for management. Virtualization is the principle technology for
access to services are factors to be focused. A good task Cloud computing. It provides the agility in operations, and
scheduler should adapt its scheduling strategy to the changing reduces cost by increasing infrastructure utilization. On the
environment and load balancing Cloud task scheduling policy.
other hand, Cloud computing automates the process through
Therefore, in this paper, Artificial Bee Colony (ABC) is applied
to optimize the scheduling of Virtual Machine (VM) on Cloud which the users can utilize resources on-demand.
computing. The main contribution of work is to analyze the Virtualization [3], [4] is critical to Cloud computing
difference of VM load balancing algorithm and to reduce the because it delivers services by providing a platform for
makespan of data processing time. The scheduling strategy was optimizing complex Information Technology resources in a
simulated using CloudSim tools. Experimental results scalable manner. The main aim of the virtualization is an
indicated that the combination of the proposed ABC algorithm, ability to run the multiple VMs on a single machine by
scheduling based on the size of tasks, and the Longest Job First
(LJF) scheduling algorithm performed a good performance sharing all the resources that belong to the hardware.
scheduling strategy in changing environment and balancing The problem of load balancing services occurs when the
work load which can reduce the makespan of data processing services from the users make a request to access to the same
time. server while other servers have no request from the services.
This phenomenon is called distributed load imbalance
Index Terms—Artificial Bee Colony, Cloud Computing, system. In this situation, it can be solved by scheduling the
Scheduling Management, Virtualization Machine
tasks or the services before using the system. Therefore, a
good task scheduler can increase the performance of
resource utilization and can reduce the makespan of
I. INTRODUCTION
assigned tasks which is called distributed load balance
ECENTLY, internet communication has become one of
R the most important factors in daily life activities.
Consequently, internet is the center of data sharing. Due to
system.
Task scheduling in distribution system such as Cloud is
used for balancing work load. It requires some conditions
the large amount of increasing users, the data is often for example, stability of the system, makespan of work,
changed or modified. This causes the heavy task scheduling ability to adapt to the environment changing, and etc. Those
for computers while less services can be processed to serve conditions in task management are similar to the behaviors
numerous users. However, this problem can be solved by of Artificial Bee Colony algorithm (ABC algorithm). ABC
using new technology of Cloud computing which focuses on is inspired by the behavior of a honey bee colony in nectar
networking and computing, storage, and data service collection. Bee colony based on the social behavior has the
resources. ability to adapt to the environment changing.
Cloud computing is an emerging concept of information The suggestion for enhancing the performance of cloud
technology service. It consists of both infrastructure as well task scheduling is to reduce the makespan of a given task set
as the application services and focuses on types of users’ and increate performance of resource utilization by using
requirements. The users can identify their needs through the improved ABC algorithm. It can be used to solve the
software in Cloud. problem and the appropriate value can be found. Moreover,
Cloud computing [1], [2] is a terminology used in it has the ability to adapt to the environment changing [5]-
describing various computing concepts that involve a large [7].
number of computers connected through a real-time The structure of this paper is organized as follows: the
communication network such as the internet. Cloud section II presents an overview of the related work. The
computing is the result of evolution and adoption of existing
Artificial Bee Colony is presented in section III. Section IV
technologies and paradigms. Its goal is to allow the users to
presents Artificial Bee Colony inspired scheduling and load
Manuscript received November 21, 2013; revised December 3, 2013.
balancing algorithm. Section V is the experimental results
B. Kruekaew is with the Department of Computer Science, Faculty of and discussion. Finally the conclusion and future work are
Science, King Mongkut’s Institute of Technology Ladkrabang (KMITL), presented in section VI.
Chalongkrung Rd., Ladkrabang, Bangkok, 10520, Thailand (e-mail:
[email protected]).
W. Kimpan is with the Department of Computer Science, Faculty of
Science, King Mongkut’s Institute of Technology Ladkrabang (KMITL),
Chalongkrung Rd., Ladkrabang, Bangkok, 10520, Thailand (e-mail:
[email protected]).

ISBN: 978-988-19252-5-1 IMECS 2014

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the International MultiConference of Engineers and Computer Scientists 2014 Vol I,
IMECS 2014, March 12 - 14, 2014, Hong Kong

II. RELATED WORK Mizan et al. [19] presented job scheduling in Hybrid
Scheduling and load balancing techniques are crucial for cloud by modifying Bee Life algorithm and Greedy
implementing the efficient parallel and distributed algorithm to achieve an affirmative response from the end
applications in making the appropriate use of parallel and users and utilize the resources.
distributed system. Task scheduling can increase the
efficiency of Cloud computing if it is able to work under the III. ARTIFICIAL BEE COLONY ALGORITHM
reasonable resources [4], [8], [9]. Artificial Bee Colony (ABC) algorithm was proposed by
Li et al. [10] proposed the tree structure method in order Karaboga [5], [6], [20], [21]. It is the method to find the
to solve Job-Shop Scheduling in supporting the needs of a appropriate value. This method is inspired by the foraging
complex and flexibility to Job-Shop Scheduling problem. It behavior of honey bees. In ABC model, there are three kinds
also appropriately managed the specific features of model of honey bee to search food sources, which include scout
management and scheduling algorithm by using Reusable bees search for food source randomly, employed bees search
Scheme and Multi-Agent system. Later, Fang et al. [11] around the food source and share food information to the
presented task scheduling in Cloud computing considering onlooker bees, and onlooker bees calculate the fitness and
the requirements of the users and load balance in Cloud select the best food source. In the nature, bees can extend
environment when the users increase and change the themselves over long distances in multiple directions. After
environment. The scheduling strategy was simulated by scout bees find the food source and return to the hive, they
using CloudSim toolkit package. compare the quality of food source and go to the dance floor
Calheiros et al. [12], [13] developed the CloudSim to perform a dance known as “waggle dance”. The waggle
simulation for modeling and simulation of virtualized dance is the communication of bees to shares the
Cloud-based datacenter environments, including dedicated information about direction of the food source, distance
management interface for VMs, memory, storage, and from the hive, and the nectar amount of the food source.
bandwidth. Simulation-based approaches can evaluate While sharing information, bees evaluate the nectar quality
Cloud computing system and application behaviors offering and energy waste. After sharing information on the dance
significant benefits. floor, onlooker bees select the best food source and then
Mondal et al. [14] proposed troubleshoot of load balance scout bees will return to the food source to bring nectar back
in Cloud computing using Stochastic Hill Climbing. In to the hive. A pseudo code of ABC algorithm is described in
2011, Hsu et al. [9] presented a focus on energy efficiency Fig. 1.
in datacenter by considering the task scheduling to physical
server and reducing energy in the system. The criterion used
to measure the performance of an energy loss was to prepare
computer exceeds the requirement of the job. Experiment
strategies for allocating server to a sequence of jobs were a
largest machine first heuristic, a best fit method, and a
mixed method. The experiment results indicated that all
three algorithms waste less energy consumption in over-
provision.
In 2010, Hu et al. [15] introduced the scheduling strategy
on load balancing of VM resource in Cloud computing
environment by adopting a tree structure using Genetic
algorithm for scheduling. It considered previous data and the
current state of work in advance to the performance
behavior of the system which can solve the problem of load
imbalance in Cloud computing.
In 2012, Wei et al. [16] presented Genetic algorithm for
scheduling in Cloud computing to increase the system
performance. Li et al. [17] proposed a Load Balancing Ant
Fig. 1. Basic of ABC algorithm.
Colony Optimization (LBACO) Algorithm. ACO was use to
schedule on load balancing and reduce makespan in Cloud. IV. ARTIFICIAL BEE COLONY INSPIRED SCHEDULING AND
Karaboga et al., in 2012 [7], presented ABC algorithm LOAD BALANCING ALGORITHM
which was used to solve the problem and find the most
appropriate value within environment changing. ABC Cloud computing provides a dynamic resource pool of
algorithm based on the swarm-based optimization algorithm VM according to different requirements from the users or
and it is used to solve the optimization problems in the system. The routing of services which request to the
searching, electrical engineering, data mining, and etc. In diverse servers, depend on the Cloud management policies
the same year, Bitam et al. [18] proposed Bee Life based on load of individual server. Scheduling management
algorithm which was used for scheduling in Cloud on Cloud computing is different degrees of loads on each
computing. Bee Life algorithm is inspired by the behavior and every VM. This may lead to different load among the
and reproduction of bee to find food source. The algorithm VMs. Load balancing techniques are effective in reducing
evaluated the performance of the resources and it has the the makespan and response time.
aim to reduce time and complexity of work.

ISBN: 978-988-19252-5-1 IMECS 2014

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the International MultiConference of Engineers and Computer Scientists 2014 Vol I,
IMECS 2014, March 12 - 14, 2014, Hong Kong

Let 𝑉𝑀 = {𝑉𝑀1 , 𝑉𝑀2 , 𝑉𝑀3 , … , 𝑉𝑀𝑁 } is a set of 𝑁 virtual 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑗 = 𝑝𝑒_𝑛𝑢𝑚𝑗 × 𝑝𝑒_𝑚𝑖𝑝𝑠𝑗 + 𝑣𝑚_𝑏𝑤𝑗 (2)
machines and 𝑇𝑎𝑠𝑘 = {𝑡𝑎𝑠𝑘1 , 𝑡𝑎𝑠𝑘2 , 𝑡𝑎𝑠𝑘3 , … , 𝑡𝑎𝑠𝑘𝐾 } is a
set of 𝐾 task to be scheduled and processed in 𝑉𝑀. All the Where 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑗 is a capacity of 𝑉𝑀𝑗 , 𝑝𝑒_𝑛𝑢𝑚𝑗
machines are unrelated but are paralleled. Scheduling is
(processing element) is the number processor in 𝑉𝑀𝑗 ,
non-preemptive which means that the processing of the
𝑝𝑒_𝑚𝑖𝑝𝑠𝑗 is million instructions per second of each
tasks on VMs cannot be interrupted. The flowchart of the
VM scheduling and load balancing using ABC algorithm is processor in 𝑉𝑀𝑗 , and 𝑣𝑚_𝑏𝑤𝑗 is the network bandwidth
shown in Fig. 2. ability of 𝑉𝑀𝑗 .

C. Select m Sites for Neighborhood Search

Scout bees with the highest fitness are chosen as
“Selected Bee” and the sites visiting by them are chosen
from neighborhood of m VMs.

D. Recruit Bees for Selected Site

Send more bees to neighborhood of the best e VM, then
evaluate the fitness based on (3).

∑𝑛 𝑡𝑎𝑠𝑘_𝑙𝑒𝑛𝑔𝑡ℎ + 𝐼𝑛𝑝𝑢𝑡𝐹𝑖𝑙𝑒_𝑙𝑒𝑛𝑔𝑡ℎ
𝑖
𝑓𝑖𝑡𝑖𝑗 = 𝐸𝑣𝑎𝑙𝑢𝑎𝑡𝑒
𝑖=1
𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝑜𝑓 𝑉𝑀 (𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑗)
(3)
𝑗

Where 𝐼𝑛𝑝𝑢𝑡𝐹𝑖𝑙𝑒_𝑙𝑒𝑛𝑔𝑡ℎ is the length of the task before

execution.

E. Select the Best Fitness Bees from Each Patch and

Assign Task to 𝑉𝑀𝑗
For each round, the bee with the best fitness will be
chosen to assign task in 𝑉𝑀𝑗 .

F. Calculate Load Balance

After submitting tasks to the under loaded 𝑉𝑀𝑗 , the
current workload of all available VMs can be calculated by
using the information that received from the datacenter.
Thus, Standard Deviation (S.D.) is calculated in order to
measure the deviations of load on VMs. Standard deviation
of load can be defined as (4):

1
𝑆. 𝐷. = �𝑛 ∑𝑛𝑗=0(𝑋𝑗 − 𝑋�)2 (4)

Fig. 2. Flowchart of the VM scheduling and load balancing using ABC

Processing time of a VM:
algorithm.

∑𝑘
𝑖=1 𝑡𝑎𝑠𝑘_𝑙𝑒𝑛𝑔𝑡ℎ𝑖
A. Initialize Population 𝑋𝑗 = 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑗
(5)
At the beginning, the initial 𝑛 scout bees are placed
randomly in 𝑉𝑀𝑠 on Cloud computing and 𝑛 is the number Mean of processing time of all VMs:
of scout bees.
∑𝑛
𝑗=1 𝑋𝑗
B. Evaluate the Fitness of the Population 𝑋� = 𝑛
(6)
Fitness is calculated based on (1).
Where 𝑆. 𝐷. is Standard deviation of load, 𝑛 is number of
∑𝑛 all VM. 𝑋𝑗 is processing time of 𝑉𝑀𝑗 which can be
𝑖=1 𝑡𝑎𝑠𝑘_𝑙𝑒𝑛𝑔𝑡ℎ𝑖𝑗
𝑓𝑖𝑡𝑖𝑗 = 𝐸𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝑜𝑓 𝑉𝑀 (𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑗)
(1) calculated based on (5) and 𝑋� is mean processing time of
𝑗
N virtual machines which can be calculated based on (6).
Where 𝑓𝑖𝑡𝑖𝑗 is the fitness of the bees population of 𝑖 in If the S.D. of the loaded VM is less than or equal to the
mean, then the system is in a balance state. On the other
𝑉𝑀𝑗 or capacity of 𝑉𝑀𝑗 with bee number of 𝑖. 𝑡𝑎𝑠𝑘_𝑙𝑒𝑛𝑔𝑡ℎ
hand, if the S.D. is higher than the mean, then the system is
is the length of the task that has been submitted in 𝑉𝑀𝑗 and in an imbalance state.
𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑗 is the capacity of 𝑉𝑀𝑗 calculating based on (2).

ISBN: 978-988-19252-5-1 IMECS 2014

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the International MultiConference of Engineers and Computer Scientists 2014 Vol I,
IMECS 2014, March 12 - 14, 2014, Hong Kong

V. EXPERIMENTAL RESULTS There are two experiments; the first experiment is the
comparison of the average makespan with the number of
A. Implementation Environment
requests changing, the second phase is the comparison of the
According to the algorithm described above, the average makespan with the number of VMs changing.
simulation using CloudSim-3.0.1 Tools will be addressed. In the first experiment, the comparison of the average
The experiments consist of 10 datacenters and 100-700 tasks makespan with the number of requests changing from the
under the simulation platform. The parameters setting on the different scheduling algorithm is performed. The number of
CloudSim-3.0.1 is shown in Table I. VMs is fixed as 50 VMs and the number of requests
TABLE I gradually increased between 100 – 700 tasks. The x axis
PARAMETERS SETTING OF CLOUD SIMULATOR shows the effect on performance of increased requests as
Type Parameter Value shown in Fig. 3.
Datacenter Number of Datacenter 10
Number of Host 5
Type of Manager Space_shared,
Time_shared
Number of PE per Host 2-4
Bandwidth 2000
Host Memory (MBa) 2048 - 10240
Datacenter Cost (The cost of using 10
processing in this resource)
Virtual Total number of VMs 30 - 210
Machine MIPSb of PEc 1000 - 2000
(VM)
Number of PE per VM 1
VM Memory (MB) 512 - 2048
Bandwidth (Bit) 1000
Type of Manager Time_shared
Task Total number of Tasks 100 - 700
Length of Task (MId) 5000 - 20000
Number of PE per requirement 1 Fig. 3. Comparison of average makespan among FCFS, LJF, SJF,
Type of Manager Space_shared ABC_FCFS, ABC_LJF, and ABC_SJF algorithm based on the fixed
Number of VMs and the increased number of the requests.
a
MB = Megabyte, bPE = Processing Element, cMIPS (Million
Instructions Per Second) is a measure of the processing speed of the
computer, MId is Million Instruction. Fig. 3 shows the average makespan of the scheduling
using improved ABC algorithm with FCFS, LJF, and SJF
B. Parameters Setting of ABC Algorithm (ABC_FCFS, ABC_LJF, and ABC_SJF), comparing with
The parameters setting of improved ABC algorithm is the original scheduling (FCFS, LJF, and SJF) which was
shown in Table II. based on 50 fixed VMs and the number of increasing
requests. The experimental results indicated that, while the
TABLE II
number of requests was increasing, the average makespan
PARAMETERS SETTING OF IMPROVED ABC ALGORITHM
consequently increased. It can be concluded that ABC_LJF
Symbol Parameter Value performed the effective results in the optimal scheduling on
n Number of scout bees 1000 the loaded system.
m Number of sites selected out of n 5
visited site In the second experiment, the comparison of the average
e Number of best site out of m select site 1 makespan with the number of VMs changing from the
nep Number of bees recruited for best e site 800 different scheduling algorithm is performed. The number of
nsp Number of bees recruited for other 200 tasks was fixed as 500 tasks and gradually increased the
(m-e) selected sites number of VMs from 30 to 210 VMs. The x axis shows the
effect on performance in increasing the system size (VMs).
C. Experimental Results Fig. 4 shows the average makespan of the scheduling
The evaluation of the performance of the proposed using ABC_FCFS, ABC_LJF, and ABC_SJF comparing
method can be described and compared with the scheduling with FCFS, LJF, and SJF which was based on the fixed
algorithm as the followings: number of tasks and the number of increasing VMs. It can
- First Come First Serve (FCFS) is considering the be concluded that ABC_LJF performed better than all
sequence of arrival of the job. methods and its performance is more prominent in
- Shortest Job First (SJF) is considering the size of the scalability.
job by selecting the shortest job first.
- Longest Job First (LJF) is considering the size of the
job by selecting the longest job first.

ISBN: 978-988-19252-5-1 IMECS 2014

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the International MultiConference of Engineers and Computer Scientists 2014 Vol I,
IMECS 2014, March 12 - 14, 2014, Hong Kong

applications,” in Artificial Intelligence Review 2012, Springer

Science Business Media B.V. 2012, March 2012.
[8] S. T. Maguluri, R. Srikant, and L. Ying, “Stochastic models of load
balancing and scheduling in cloud computing clusters,” in Proc.
IEEE Infocom, 2012, pp. 702-710.
[9] Y. C. Hsu, P. Liu, and J. J. Wu, “Job sequence scheduling for cloud
computing,” in Int. Conf. on Cloud and Service Computing (CSC
2011), 2011, Dec. 2011, pp. 212-219.
[10] N. Li, Z. Hong, and H. Xiaoting, “Dynamic integration mechanism
for job-shop scheduling model base using Multi-agent,” in 2009 Int.
Conf. on Information Management, Innovation Management and
Industrial Engineering, 2009, vol. 4, Dec. 2009, pp. 421-425.
[11] Y. Fang, F. Wang, and J. Ge, “A task scheduling algorithm based on
load balancing in cloud computing,” in Web Information Systems and
Mining, Lecture Notes in Computer Science 2010, Vol. 6318, 2010,
pp. 271-277.
[12] R. N. Calheiros, R.Ranjan, C. A. F. D. Rose, and R. Buyya,
“CloudSim: a toolkit for modeling and simulation of cloud computing
environments and evaluation of resource provisioning algorithms,”
Software : Practice and Experience, Vol. 41, No.1, Jan. 2011, pp. 23-
50.
Fig. 4. Comparison of average makespan between FCFS, LJF, SJF, [13] R. N. Calheiros, R.Ranjan, C. A. F. D. Rose, and R. Buyya,
ABC_FCFS, ABC_LJF, and ABC_SJF algorithm based on the fixed “CloudSim: A novel framework for modeling and simulation of cloud
number of the requests and the increased number of VMs. computing infrastructures and services,” arXiv preprint arXiv:
0903.2525, 2009.
[14] B. Mondal, K. Dasgupta, and P. Dutta, “Load balancing in cloud
VI. CONCLUSION AND FUTURE WORK computing using Stochastic Hill Climbimg-A soft computing
This paper presents ABC optimization algorithm which approach,” in Procedia Tachnology, vol. 4, 2011, pp. 783-789.
[15] J. Hu, J. Gu, G. Sun, and T. Zhao, “A scheduling strategy on load
can solve the Virtual machine scheduling management balancing of virtual machine resources in cloud computing
under the environmental changing of the number of VMs environment,” in 3rd Int. Symp. on Parallel Architectures,
and requests on Cloud computing. Even in changing Algorithms and Programming(PAAP), 2010, 18-20 Dec. 2010, pp. 89-
96.
environment, Cloud computing needs to be operated in a [16] Y. Wei, and L. Tian. “Research on cloud design resources scheduling
stable system. Therefore, ABC algorithm is suitable for based on genetic algorithm,” in 2012 Int. Conf. on Systems and
Cloud computing environment because the algorithm is able Informatics (ICSAI 2012), 2012, May 2012, pp. 2651-2656.
[17] K. Li, G. Xu, G. Zhao, Y. Dong, and D. Wang, “Cloud task
to effectively utilize the increased system resources and scheduling based on load balancing Ant Colony Optimization,” in 6th
reduce makespan. The experimental results illustrated that Annual ChinaGrid Conf., 2011, Aug. 2011, pp. 3-9.
the proposed methods of ABC_LJF performed effective [18] S. Bitam, “Bees life algorithm for job scheduling in cloud
computing,” in Conf. on Computing and Information Technology
results than all methods and its performance is more (ICCIT 2012), 2012, pp. 186-191.
prominent in scalability. Under the circumstance of [19] T. Mizan, S. M. R. A. Masud, and R. Latip, “Modified bees life
increasing or decreasing the number of servers, the load algorithm for job scheduling in hybrid cloud,” in Int. Journal of
Engineering and Technology(IJET), 2012, vol. 2, no.6, June 2012, pp.
balancing algorithm should be done by using ABC_LJF
974-979.
algorithm in order to maintain system stability and [20] D. Karaboga, “An Idea based on honey bee swarm for numerical
scheduling and to prevent the system crash. For further optimization,” in Technical Report TR06, Erciyes University,
studies, the preemptive Virtual machine scheduler operating Engineering Faculty, Computer Engineering Department, 2005.
[21] D.T. Pham, A. Ghanbarzadeh, E. Koc, S. Otri, S. Rahim, and M.
with independent and heterogeneous tasks on Cloud Zaidi, "The Bees Algorithm-A novel tool for complex optimisation
computing will be focused. problems," in Proc. of IPROMS 2006 Conf., 2006, pp. 454-461.

REFERENCES
[1] K. Tang, J. M. Zhang, and C. H. Feng, “Application centric lifecycle
framework in cloud,” IEEE 8th Int. Conf. on e-Business Engineering
(ICEBE), 2011, 19-21 Oct. 2011, pp. 329-334.
[2] M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A.
Konwinski, G. Lee, D. Patterson, A. Rabkin, I. Stoica, and M.
Zaharia, “A view of cloud computing,” in Communications of the
acm, vol. 53, no. 4, April 2010, pp. 50-58.
[3] K. Dutta, “A smart job scheduling system for cloud computing
service providers and users: modeling and simulation,” in 1st Int’l
Conf. on Recent Advances in Information Technology RAIT-2012,
2012.
[4] S.Sindhu, and S. Mukherjee, “Efficient task scheduling algorithms for
cloud computing environment,” in High Performance Architecture
and Grid Computing, Communications in Computer and Information
Science, 2011, vol. 169, 2011, pp. 79-83.
[5] D. Karaboga, and B. Basturk, “Artificial Bee Colony (ABC)
optimization algorithm for solving constrained optimization
problems,” in IFSA 2007, LNAI 4529, 2007, pp. 789–798.
[6] D. Karaboga, and B. Basturk, “On the performance of artificial bee
colony (ABC) algorithm,” in Applied Soft Computing, 2008, pp. 687–
697.
[7] D. Karaboga, B. Gorkemli, C. Ozturk, and N. Karaboga, “A
comprehensive survey: artificial bee colony (ABC) algorithm and

ISBN: 978-988-19252-5-1 IMECS 2014

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)