Ieeepro techno solutions 2014 ieee dotnet project - a hyper-heuristic scheduling

A Hyper-Heuristic Scheduling
Algorithm for Cloud
Chun-Wei Tsai, Wei-Cheng Huang, Meng-Hsiu Chiang, Ming-Chao Chiang, and Chu-Sing Yang
Abstract—Rule-based scheduling algorithms have been widely used on many cloud computing systems because they are simple and
easy to implement. However, there is plenty of room to improve the performance of these algorithms, especially by using heuristic
scheduling. As such, this paper presents a novel heuristic scheduling algorithm, called hyper-heuristic scheduling algorithm (HHSA), to
find better scheduling solutions for cloud computing systems. The diversity detection and improvement detection operators are
employed by the proposed algorithm to dynamically determine which low-level heuristic is to be used in finding better candidate
solutions. To evaluate the performance of the proposed method, this study compares the proposed method with several state-of-the-art
scheduling algorithms, by having all of them implemented on CloudSim (a simulator) and Hadoop (a real system). The results show that
HHSA can significantly reduce the makespan of task scheduling compared with the other scheduling algorithms evaluated in this
paper, on both CloudSim and Hadoop.
Index Terms—Cloud computing, evolutionary algorithm, scheduling, and Hadoop
Ç
1 INTRODUCTION
UNTIL now we are still actively looking for possible
solutions to enhance the performance of information
systems for computation, analysis, and storage. Since
distributed and parallel computing was widely used to
enhance the performance of a variety of computer systems,
several strategies have been presented for different
approaches and congenital restrictions in different eras. No
matter which consideration it is for, how to efficiently man-
age computer resources is definitively a vital research issue.
Among them, scheduling [1], [2], [3] is essential in the suc-
cess of enhancing the performance of a computer system.
With the advance of computer and internet technologies,
paradigms of cloud computing [4], [5], [6], [7] have been
successfully used on several information systems in recent
years. Although cloud computing systems nowadays pro-
vide a better way to carry out the submitted tasks in terms
of responsiveness, scalability, and flexibility, most job and
task scheduling problems on cloud computing systems are
still either NP-hard or NP-complete [8]. According to our
observations, most rule-based scheduling algorithms (e.g.,
exhaustive and deterministic algorithms) are widely used
on today’s cloud computing systems because they are sim-
ple and easy to implement. Unfortunately, these rule-based
scheduling algorithms are inappropriate for large-scale or
complex scheduling problems because the results of these
scheduling strategies are usually far from optimal. This
implies that there is plenty of room to improve the schedul-
ing algorithms for cloud computing systems.
A promising research direction, which applies modern
heuristics [9] to scheduling on cloud computing systems,
has attracted several researchers from different research
domains. The extended version of heuristics combines two
or more heuristics into a single heuristic algorithm to lever-
age their strengths for scheduling, called hybrid heuristic-
based scheduling algorithm. Different from most hybrid
heuristic-based scheduling algorithms that employed only
one or two heuristics, the proposed algorithm is a hyper-
heuristic-based algorithm that integrates several heuristic
algorithms. The basic idea of the proposed algorithm is to
leverage the strengths of heuristic algorithms, such as simu-
lated annealing [10], genetic algorithm [11], particle swarm
optimization [12], and ant colony optimization [13], by inte-
grating them into a single algorithm. The main contribu-
tions of the paper can be summarized as follows:
1) A novel high-performance hyper-heuristic algorithm
is proposed for scheduling on cloud computing sys-
tems to reduce the makespan.
2) Two detection operators—one for diversity detection
and one for improvement detection—are proposed
for the proposed algorithm to determine the timing
to employ the low-level heuristic (LLH) algorithm.
3) The proposed algorithm can be applied to both
sequence-dependent (see also Section 5.1.1) and
sequence-independent (see also Section 5.2.1) sched-
uling problems.
4) To evaluate its performance on both a cloud simula-
tor and a real cloud system, the proposed algorithm
and several other scheduling algorithms (for the pur-
pose of comparison) are implemented on both the
CloudSim [14] and Hadoop.
C.W. Tsai is with the Department of Applied Informatics and Multimedia,
Chia Nan University of Pharmacy Science, Tainan 71710, Taiwan.
E-mail: cwtsai0807@gmail.com.
W.C. Huang, M.H. Chiang, and M.C. Chiang are with the Department of
Computer Science and Engineering, National Sun Yat-sen University,
Kaohsiung 80424, Taiwan. E-mail: demon2506@yahoo.com.tw,
m993040011@student.nsysu.edu.tw, mcchiang@cse.nsysu.edu.tw.
C.S. Yang is with the Department of Electrical Engineering, National
Cheng Kung University, Tainan 70101, Taiwan.
E-mail: csyang.ee.ncku.edu.tw.
Manuscript received 15 Sept. 2013; revised; 16 Mar. 2014; accepted 18 Mar.
2014. Date of publication 6 Apr. 2014; date of current version 30 July 2014.
Recommended for acceptance by I. Bojanova, R.C.H. Hua, O. Rana, and M.
Parashar.
For information on obtaining reprints of this article, please send e-mail to:
reprints@ieee.org, and reference the Digital Object Identifier below.
Digital Object Identifier no. 10.1109/TCC.2014.2315797
236 IEEE TRANSACTIONS ON CLOUD COMPUTING, VOL. 2, NO. 2, APRIL-JUNE 2014
2168-7161 ß 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

5) A detailed analysis of the parameter settings is also
given to show the impact they may have on the per-
formance of the proposed algorithm.
The remainder of the paper is organized as follows.
Section 2 begins with a brief introduction to the traditional
scheduling, followed by the scheduling on cloud comput-
ing. In addition to the scheduling problem and its relevant
technologies, Section 3 provides a brief discussion on heu-
ristic algorithms and compares the differences between
hybrid-heuristic and hyper-heuristic algorithms. Section 4
describes in detail the proposed algorithm and gives a sim-
ple example. The simulation results, on both CloudSim and
Hadoop, are discussed in Section 5. Conclusions and future
trends are drawn in Section 6.
2 SCHEDULING
2.1 Scheduling Problems
2.1.1 Traditional Scheduling
The scheduling problem [15], [16] can be defined as follows:
Find an optimal solution to schedule a given set of tasks
T ¼ fT1; T2; . . . ; Tng to a given set of machines M ¼ fM1;
M2; . . . ; Mmg subject to a predefined set of constraints and
measurements. For instance, one of the widely used meas-
urements for the scheduling problem is the so-called make-
span CmaxðsÞ, which is defined as the completion time of the
last task. Mathematically, the scheduling problem is to
minimize fðsÞ ¼ CmaxðsÞ; (1)
where s is a candidate solution, and, letting Cj denote the
completion time of job j, CmaxðsÞ ¼ maxj Cj is the comple-
tion time of the last job. Moreover, with advance of com-
puter systems, the scheduling problem has evolved from
single processor, single tasking, to multiprocessor, multi-
tasking, then to large scale distributed system nowadays.
To measure the solutions for different kinds of scheduling
problems, makespan is no longer the only choice. Instead, a
number of measurements have been developed. Examples
are maximum lateness, maximum tardiness, and maximum
flowtime [1].
Several traditional scheduling algorithms have been pre-
sented, such as earliest due date first (EDD) [17], critical
path method (CPM) [18], project evaluation and review
technique (PRET) [19], dynamic programming [20], and
branch-and-bound [21]. These methods are generally much
easier to implement than metaheuristic and hybrid schedul-
ing algorithms because they are all designed based on one
or a few particular rules (ways) to manage and arrange the
tasks, but metaheuristic and hybrid scheduling algorithms
are not. Some of them even guarantee to find the optimal
solution for the scheduling problem in question, but they
are all suffered from the fact of not being able to find the
optimal solution in a reasonable time, especially when the
problem becomes either too complex or too large.
In fact, it is the other two important disadvantages that
take the researchers into consideration of modifying the
search strategy of these traditional scheduling algorithms.
First, most full search algorithms, such as the well-known
branch-and-bound and dynamic programming, are nor-
mally time consuming because the number of checks they
have to perform is very large. Second, in the case of the
deterministic algorithms (such as EDD), although they are
very fast and easy to implement, they are easily falling into
local optima. The classical exhaustive algorithms are gener-
ally good enough to provide the optimal solution quickly
for small scheduling problems [22], but not large scheduling
problems [23]. In this paper, by the small scheduling prob-
lems, we mean problems for which all the solutions can be
checked in a reasonable time by using classical exhaustive
algorithms running on modern computer systems. In con-
trast, by the large scheduling problems, we mean problems
for which not all the solutions can be checked in a reason-
able time by using the same algorithms running on the
same computer systems. These observations make it easy to
understand that exhaustive algorithms will take a prohibi-
tive amount of time to check all the candidate solutions for
large scheduling problems because the number of candidate
solutions is simply way too large to be checked in a reason-
able time. As a result, researchers have paid their attention
to the development of scheduling algorithms that are effi-
cient and effective, such as heuristics.
2.1.2 Scheduling on Cloud
With computing systems being shifted to cloud-based sys-
tems progressively, one of the main characteristics is that it
works on a pay-as-you-use basis [24]. Several studies
attempted to define the scheduling problem on cloud sys-
tems as the workflow problem, which can be further classi-
fied into two levels: service-level (platform layer and static
scheduling) and task-level (unified resource layer and
dynamic scheduling) [25]. Different from grid computing,
the user can install their programs on the virtual machines
(VMs) and determine how to execute their programs on the
cloud computing system. For these reasons, although both
grid computing and cloud computing are heterogeneous,
the key issues they face are very different. A good example
is the cost and latency of data transfer on these environ-
ments. That is why some studies added more considerations
to their definitions of scheduling on cloud. For instance, a
couple of studies [26], [27] used directed acyclic graph
(DAG) to define the scheduling problem on cloud. The basic
idea is to use the vertices of a DAG to represent a set of tasks
and the edges between the vertices to represent the depen-
dencies between the tasks. Then, the scheduling problem on
cloud can be formulated for the solution s as follows [27]:
minimize fðsÞ ¼ CmaxðsÞ þ
Xn
i¼1
Xm
j¼1
Cij;
subject to CmaxðsÞ UðsÞ;
CðsÞ BðsÞ;
(2)
where fðsÞ is the objective function; CmaxðsÞ the completion
time of the last job (also called makespan); n the number
of tasks; m the number of machines; Cij the cost of process-
ing the ith task on the jth machine; UðsÞ the number of over-
due tasks; CðsÞ ¼
Pm
j¼1 Cij the total cost of s; and BðsÞ the
restriction on the budget for the tasks of s. Moreover, since
the study described in [27] also takes into account the maxi-
mum expenditure for the provisioning of services or appli-
cations on cloud, such as QoS, the budget is also considered
as part of the scheduling problem for cloud when necessary.
TSAI ET AL.: A HYPER-HEURISTIC SCHEDULING ALGORITHM FOR CLOUD 237

Recently, Hadoop [28] was widely used in deploying
cloud computing system. In addition to being used by sev-
eral international companies (e.g., IBM, Google, AOL, eBay,
Yahoo, and Amazon) [29], the features provided by the
open-source environment and MapReduce framework have
made it a very popular cloud computing platform [30]. Its
scalability for large-scale data and systems has also been
found in recent researches, such as big data approach [31]
and smart grid [32]. The basic idea is so that the users of
such a system can submit jobs to the Hadoop cluster, which
will then split the jobs into tasks of a given size and provide
a number of slots to run tasks in parallel. The processing
speed of these slots depends on the types of jobs submitted
to the cluster. The definitions of parallel machines for sched-
uling can be divided into three kinds: (1) identical machines
in parallel, (2) machines in parallel with different speeds,
and (3) unrelated machines in parallel [1]. Since a Hadoop
cluster generally consist of machines with different speeds,
its task scheduling problem can be considered as the second
kind of parallel machines for scheduling.
2.2 Scheduling Methods for Cloud
To develop more efficient scheduling algorithms for cloud,
some recent reviews were given in [33], [34], which encom-
pass non-metaheuristic and metaheuristic scheduling
algorithms, such as earliest-finish-time-based algorithm and
genetic algorithm. Also given in [33] was a comparison
sheet to classify the scheduling algorithms discussed there
by methods, parameters, and factors of scheduling. In addi-
tion, the method presented in [35] uses feedback informa-
tion to estimate the earliest finish time to dynamically
adjust the resource allocation.
Among researches on non-metaheuristic scheduling
algorithms, the earliest start time, latest start time, and
makespan [36], [37] are usually used to evaluate the sched-
uling results; however, to precisely describe cloud environ-
ment, additional measurements (objective functions) are
also used in describing and measuring the scheduling on
cloud. For instance, the cost of workflow and the deadline
defined by a user were used in [38] while the price-based
model was used in [39]. Besides the additional measure-
ments, another study [36] combined the static and dynamic
scheduling algorithms into a cloud system to adjust its
resource allocation strategies on the fly. The main advan-
tages of this kind of algorithms are: (1) they are simple and
easy to implement; (2) some rule-based deterministic
algorithms can find acceptable solutions quickly; and (3)
most of them are compatible to each other; thus, some stud-
ies have tried to integrate two or more non-metaheuristic
algorithms to solve the scheduling problem. Theoretically,
one of the major disadvantages of these algorithms is in that
the results obtained by these algorithms may be far from
optimal or even acceptable. In practice, our observation
shows that because they are simple and easy to implement
on a cloud system, several cloud systems use traditional
scheduling algorithms to manage the computing resources,
such as Hadoop, which uses first in first out (FIFO) [40] as
the default scheduling algorithm.
Although the FIFO scheduler of Hadoop can achieve a
high utilization as far as a cluster is concerned, it provides
no fairness at all. Even worse, it sometimes leads to a long
response time, especially when the number of jobs to be run
is beyond the available computing resources. For a concrete
example, consider the following scenario: A very “long” job
arrives the job queue followed by a “short” job with the
same priority. The FIFO scheduler of Hadoop will schedule
the first job and allocate all the resources it needs. This
implies that the second job must wait until the first job is
completed. The result is obviously a very long response
time for the second job. For a cluster that has enough capac-
ity to run all the jobs at the same time, most schedulers can
satisfy the demand of the input jobs and tasks. It means that
each job can be run and completed before the deadline.
Assume a high map-capacity Hadoop cluster has 200 maps.
Most of the schedulers can finish the jobs when there are 10
input jobs each of which use 20 maps, i.e., when there are
total 10 Â 20 ¼ 200 tasks to be assigned to the 200 map slots.
When the input jobs and tasks are too large to be run by the
cluster (system) at the same time, the overall computation
time will differ because different scheduling strategies may
decide to schedule the jobs to be run and the resources to be
allocated differently. Moreover, most, if not all, of the
schedulers shipped with Hadoop today do not take into
consideration the so-called makespan—the objective hyper-
heuristic scheduling algorithm (HHSA) described herein is
trying to improve.
To solve this problem, Facebook developed Hadoop Fair
Scheduler (HFS) and improved it with Delay Scheduling
[41] to get better data locality and throughput. HFS calcu-
lates each task’s “fair share,” sorts them in increasing order,
and then picks the one with least fair share when there are
empty slots. In the scenario above, HFS will schedule the
shorter job first when it receives low fair share, thus finish-
ing the shorter job before the longer one completes, by
sacrificing only a little delay for a dramatically faster
response time of the second job. On the other hand, Yahoo
developed Capacity Scheduler [42] to deal with the same
scheduling problem. By guaranteeing the capacity of the
queues for different jobs, Capacity Scheduler can also
achieve fairness, but it requires a more sophisticated control
over the cluster. In summary, these two schedulers are not
focused on pursuing the optimal solution; rather, they try
to enhance the response time of each job. As such, this
paper is aimed at demonstrating that by estimating and
predicting the processing speed and completion time of
each machine and each job, HHSA can generate feasible sol-
utions that optimize the makespan not taken into consider-
ation by the other schedulers. In addition, our experiments
show that HHSA can eventually provide better solutions in
terms of makespan.
In addition to non-metaheuristic scheduling algorithms,
an alternative is the heuristic scheduling algorithms [9] that
can be used to provide better scheduling plans than do the
rule-based scheduling algorithms for cloud computing. As
shown in Fig. 1a, the basic idea of heuristics is to use three
key operators—transition, evaluation, and determination—to
“search” for the possible solutions on the convergence pro-
cess. Note that in Fig. 1a, t denotes the iteration number;
tmax the maximum number of iterations or the stop criteria.
More precisely, the transition operator creates the solution
s, by using methods which could be either perturbative or
constructive or both [43]; the evaluation operator measures

the fitness of s by using a predefined measurement; and
then the determination operator determines the next search
directions based on the s from the transition operator and
the evaluation operator. Because heuristics use “tactical
guess” to find the possible solutions, it has a much better
chance to find a better result than do most rule-based deter-
ministic algorithms.
3 INTEGRATION OF HEURISTIC ALGORITHMS
As we mentioned in Section 2.2, heuristic algorithms are
able to provide better results than rule-based algorithms for
complex scheduling problems on cloud computing systems.
Moreover, as far as this paper is concerned, by the complex
scheduling problems, we mean scheduling problems that
are NP-complete [8], thus having a huge solution space for
which all the solutions cannot be checked by any known
exhaustive algorithm in a reasonable time, not only in a tra-
ditional information system (non-cloud virtualized cluster)
but also in a cloud system. Each heuristic algorithm, how-
ever, has its pros and cons as far as scheduling is concerned.
For example, ant colony optimization may provide a better
scheduling solution than other scheduling algorithms, but it
takes much more computation time than the others. An
alternative to create a scheduling algorithm is by integrating
two or more heuristic algorithms or combining heuristic
algorithms with rule-based algorithms, Generally speaking,
an integration that invokes two or more heuristic algorithms
at each iteration reflects the complementary advantages of
these algorithms to find a better result than a single heuristic
algorithm does. As shown in Fig. 1b, the basic idea of
hybrid-heuristic algorithm is to combine heuristic algo-
rithms to perform the transition (T), evaluation (E), and
determination (D) at each iteration, where Hi denotes one of
the heuristic algorithms. This kind of integration may com-
pensate for the intrinsic weak points of specific heuristic
algorithms. For instance, the search diversity of tabu search
may be increased by integrating it with genetic algorithm or
other heuristics. However, a critical problem is that
although the hybrid-heuristic may have a higher chance to
find a better result than a single heuristic does, it generally
takes a longer computation time than heuristics at each iter-
ation of the convergence process.
Recently, another way to combine two or more heuristic
algorithms, called hyper-heuristic algorithm [44], has
attracted many researchers from different domains. As
shown in Fig. 1c, one of the heuristic algorithms in the pool
of candidate heuristics fH1; H2; . . . ; Hng will be selected by
the low-level heuristic selection operator as the heuristic
algorithm to be performed, which is referred to as Hi in
Fig. 1c. For this kind of heuristics, low-level heuristic selec-
tion and acceptance are the two essential operators to deter-
mine which heuristic algorithm is to be selected and
whether the selected heuristic algorithm will be continu-
ously used or not. For this reason, several LLH selection
operators have been presented in different studies [45] to
choose the Hi. Different from the LLH selection operator
that plays the role of determining the heuristic algorithm,
the acceptance operator plays the role of determining the
timing to select a new heuristic algorithm, such as when the
Hi gets stuck with a solution for a certain number of itera-
tions or when the Hi falls into local optimum.
In spite of all these extensions, there still lacks a simple
and efficient design for scheduling on cloud today. For
example, although most rule-based scheduling algorithms
are very simple, they are unable to find the optimal or
approximate solution in a reasonable time because most
scheduling problems are NP-complete. Now, let us look
back at the heuristic algorithms. Like the hybrid-heuristic,
hyper-heuristic also attempts to use two or more heuristics
on the convergence process. Unlike the hybrid-heuristic,
the basic idea of hyper-heuristic is to use “one and only
one” heuristic algorithm at each iteration. With these char-
acteristics, hyper-heuristic algorithms can then maintain a
high search diversity to increase the chance of finding bet-
ter solutions at later iterations while not increasing the
computation time.
The time complexities of heuristic, hybrid-heuristic, and
hyper-heuristic are definitely an important issue, especially
a comparison between the time complexities of these
approaches. As described in [46], the expected running time
of heuristic is OðM log MÞ where M indicates the number of
subsolutions of each solution of the problem. Tseng and
Yang in a later study [47] pointed out that the time complex-
ity of heuristic is Oð‘N M2
Þ where ‘ indicates the number of
iterations the algorithm performs; N the population size;
and M the number of subsolutions. In [48], the time com-
plexity of hybrid-heuristic algorithm for scheduling was
given. For each iteration, the computation of makespan
takes OðmnÞ, simulated annealing takes Oðmn2
Þ, and
(a) (b) (c)
Fig. 1. Outline of (a) heuristics, (b) hybrid-heuristics, and (c) hyper-heuristics.

Nawaz, Enscore, and Ham (NEH) takes Oðmn3
Þ, where m
denotes the number of machines; and n the number of jobs.
Thus, the time complexity of hybrid-heuristic algorithm is
Oð‘mn3
Þ. This is due to the fact that the NEH operator takes
Oðmn3
Þ. Putting these observations together, the time com-
plexity of hybrid-heuristic algorithm can be obtained as
Oð‘M N2
Þ, assuming that (1) m Â n of the hybrid-heuristic
algorithm equals M of the heuristic algorithm, which is usu-
ally true for the scheduling problems,1
and (2) the time
complexity of all the other operators of heuristic algorithms
are bounded from above by N M2
. Similar to the hybrid-
heuristic algorithm, the hyper-heuristic algorithm also uses
two or more heuristic algorithms, but the time complexity
of each operator is usually bounded from above by
OðM N2
Þ. As a result, the time complexity of the hyper-heu-
ristic algorithm is also Oð‘M N2
Þ if all the operators (or heu-
ristic algorithms) used are bounded from above by M N2
.
4 THE PROPOSED ALGORITHM
4.1 Hyper-Heuristic Scheduling Algorithm
The basic idea of the proposed algorithm is to use the diver-
sity detection and improvement detection operators to bal-
ance the intensification and diversification in the search of
the solutions during the convergence process. The proposed
algorithm, called Hyper-Heuristic Scheduling Algorithm, is
as shown in Fig. 2. Line 1 sets up the parameters fmax and
fni, where fmax denotes the maximum number of iterations
the selected low-level heuristic algorithm is to be run; fni
the number of iterations the solutions of the selected low-
level heuristic algorithm are not improved. Line 2 reads in
the tasks and jobs to be scheduled, i.e., the problem in ques-
tion. Line 3 initializes the population of solutions
Z ¼ fz1; z2; . . . ; zN g, where N is the population size. On line
4, a heuristic algorithm Hi is randomly selected from the
candidate pool H ¼ fH1; H2; . . . ; Hng.
As far as the proposed algorithm described herein is
concerned, the low-level heuristic candidate pool consists
of simulated annealing [10], genetic algorithm [11], parti-
cle swarm optimization [12], and ant colony optimization
[13]. The selected hyper-heuristic algorithm (LLH) Hi will
then be performed repeatedly until the termination crite-
rion is met, as shown in lines 5-13. More precisely, the
selected LLH Hi will evolve the solution Z for fmax itera-
tions by using the determine function cðHi; F1; F2Þ, as
defined in Eq. (3), to balance the intensification and diver-
sification of the search directions, which in turn rely on
the information provided by the improvement detection
operator denoted F1 (as shown in line 7) and by the diver-
sity detection operator denoted F2 (as shown in line 8) to
decide whether to select a new LLH or not. For single-
solution-based heuristic algorithms (SSBHA), only F1 is
used whereas for population-based heuristic algorithms
(PBHA), both F1 and F2 are used.
cðH; F1; F2Þ ¼
false; if H 2 S and F2 ¼ true;
false; if H 2 P and F1 ¼ true and F2 ¼ true;
true; otherwise;
8

:
(3)
where S denotes the set of SSBHAs; P the set of PBHAs.
When the determine function cðHi; F1; F2Þ returns a true,
the proposed algorithm will randomly select a new heuristic
algorithm Hi and then return the solution Z to the perturba-
tion operator to fine-tune the result, as shown in lines 9-11.
Note that the single-solution-based heuristic algorithms
employ one and only one search direction at each iteration
on the convergence process while the population-based
heuristic algorithms employ multiple search directions at
each iteration.
4.1.1 The Improvement Detection Operator
A simple random method is used to select the low-level
heuristic Hi from the candidate pool H. According to our
observation, the best so far makespan (BSFMK) for both SA
and GA could continue to improve the results at early itera-
tions (e.g., less than 200 iterations), but it is difficult to
improve the results at later iterations (e.g., after 800 itera-
tions), especially when the search directions converge to a
small number of directions, as shown in Fig. 3.
1. This means that the solution is encoded as a matrix of size m by n,
i.e., a matrix with m rows and n columns, where each row corresponds
to a machine; each column corresponds to a job. Therefore, the number
of subsolutions of each solution is m Â n ¼ M.
Fig. 2. Hyper-heuristic scheduling algorithm.
Fig. 3. Results of applying heuristic algorithms to the j30 data set for the
workflow problem on CloudSim.

From these observations, a simple approach to detecting
when to select a new LLH is as given below:
F1 ¼
false; BSFMK is not improved after tni iterations;
true; otherwise:

(4)
This approach only checks to see whether the solutions
found by Hi, i.e., BSFMK, are being improved or not. If the
selected Hi cannot improve the BSFMK after a row of fni
iterations, the improvement detection operator will return a
false value to the high level hyper center to indicate that it
should pick up a new LLH. Note that the improvement
detection operator will return a false value in three cases:
the maximum number of iterations fmax is reached, the
number of iterations fni the solutions are not improved is
reached, and when the stop condition is reached.
4.1.2 The Diversity Detection Operator
In addition to the improvement detection operator, the
diversity detection operator is used by HHSA to decide
“when” to change the low-level heuristic algorithm Hi.
Here is how it works. The diversity of the initial solution
DðZ0Þ will be used as a threshold v, i.e., v ¼ DðZ0Þ. The
diversity of the current solution DðZÞ is computed as the
average of the task distances between individual solutions
which are defined as follows: If a task in two different indi-
viduals is assigned to the same VM, the task distance is 0;
otherwise, the task distance is 1.
If the diversity of the current solution DðZÞ is less than
the threshold v (the diversity of the initial solution), this
operator will return false, and HHSA will then randomly
select a new LLH, as given below:
F2 ¼
true; DðZÞ v;
false; otherwise:

(5)
Note that v is defined as m1
À 3 Â s1
where m1
and s1
denotes, respectively, the average distance and the standard
deviation of the distances of the initial solution.
4.1.3 The Perturbation Operator
In addition to the perturbation operators of Hi itself, the
proposed algorithm will perturb the solutions obtained by
Hi before they are passed on to the newly selected LLH.
This means that the candidate solutions created by the low-
level heuristic Hi can be perturbed by the simulated anneal-
ing mechanism we presented herein, by assigning a differ-
ent temperature to each individual to balance the
intensification and diversification of the search. More pre-
cisely, the perturbation temperature Tk is used to determine
the probability of randomly changing the subsolutions,
such as changing the first subsolution of the solution 0001
from 0 to 1, thus obtaining the solution 1;001. This mecha-
nism implies that the higher the temperature, the higher the
opportunity to escape from the local search space to find
better solutions. More precisely, the perturbation tempera-
ture Tk is defined as
Tk ¼
Tmax Á
fmax À f
fmax
Á
fðzbestÞ À fðzkÞ
fðzbestÞ À fðzworstÞ
; if Hi 2 P;
Tmax Á
fmax À f
fmax
Á
1
N
; otherwise;
8

:
(6)
where Tk (1 k N) denotes the temperature assigned to
the kth individual, Tmax the maximum temperature, fmax
the maximum number of iterations LLH will be performed,
and f the number of iterations LLH is actually performed.
Also, fðÁÞ is the fitness function; zk, zbest, and zworst are,
respectively, the finesses of the kth solution, the best fitness
of the population, and the worst fitness of the population.
If the low-level heuristic is an SSBHA, it implies that no
population information will be available. In this case, the
perturbation operator will only use the number of itera-
tions to decide the temperature of the current best individ-
ual from the population of LLH which will be passed to
the newly selected LLH.
4.2 An Example
A simple example is given in Fig. 4 to depict how HHSA
works. As this example shows, there are two cases to
deal with: one is SSBHAs while the other is PBHAs. If
the randomly selected algorithm is an SSBHA, such as
simulation annealing (SA), then it will be evolved by
using the improvement detection operator to check if
there is an improvement in the fitness of the current
solution. On the contrary, if a PBHA is chosen, it is the
diversity of the current solutions (i.e., the solutions in
the population) that will be checked by using the diver-
sity detection operator before the improvement detection
operator is applied. After that, the perturbation operator
will be applied to either the current solution or the pop-
ulation before they are passed on to the newly selected
LLH. This process will be repeated until the stop condi-
tion is reached. As shown in Fig. 4, the hyper-heuristic
algorithm contains many low-level heuristic algorithms.
After the candidate low-level heuristic algorithms to be
in the pool are decided, the proposed algorithm will ran-
domly select a low-level heuristic algorithm Hi from the
pool. Then, if Hi is a PBHA, the proposed algorithm will
first perform the diversity detection operator. Finally, for
both the PBHA and SSBHA, the improvement detection
and perturbation operators will be performed. The pro-
cess will be repeated until a new low-level heuristic
algorithm is selected.
5 RESULTS
In this section, two kinds of scheduling problems are used
to evaluate the performance of the proposed algorithm. One
is the workflow scheduling problem; the other is the
Hadoop map-task scheduling problem.
Fig. 4. Example showing how HHSA works.

5.1 Workflow Scheduling on CloudSim
5.1.1 Data Sets and Parameter Settings
CloudSim [14] is a tool for emulating a cloud computing
environment. In this study, CloudSim is used as the emula-
tor in solving the workflow scheduling problem. As shown
in Fig. 5a, CloudSim can be used to construct a data center
with a set of virtual machines as the resource. Each task of a
workflow can be started by assigning it to a VM once its par-
ent tasks are completed. For example, as shown in Fig. 5b,
task 3 will be started once its parent tasks 0 and 1 are com-
pleted. Besides, the starting and ending tasks, s and e, are
virtual task and are not to be assigned to any VM. For more
details and other workflow examples, readers are referred
to [49] and [50].
The empirical analysis was conducted on an ASUS PC
with 2.67 GHz Intel i7-920 CPU and 4 GB of memory running
Fedora 12 with Linux 2.6.31 and using CloudSim to construct
four virtual machines in a data center. The workflow data
sets from the studies described in [51], [52] and PSPLIB [53]
are employed to compare the performance of the proposed
algorithm and other scheduling algorithms evaluated in this
paper. Among these data sets, the e-Economic, fMRI, and
protein annotation have, respectively, 10, 15, and 15 tasks
[51], [52] while the j30, j60, and j90 have, respectively, 30, 60,
and 90 tasks [53]. More precisely, fMRI is a data set for
evaluating workflow that can be processed in parallel while
e-Economic is a data set for evaluating workflow that can be
processed sequentially. For a scheduling problem, the data
set of [53] contains serial and parallel tasks. Each simulation
was carried out 30 runs. The maximum number of iterations
each run is set equal to 1,000. The population size is set equal
to 50 for PBHAs. In HHSA, the maximum number of itera-
tions of low-level algorithm fmax is set equal to 50, and the
non-improved iteration threshold fni is set equal to 5. The
other parameter settings of the scheduling algorithms are as
shown in Table 1.
5.1.2 Simulation Results of Workflow Scheduling
To evaluate the performance of HHSA for the workflow
scheduling problem, we compare it with two traditional
rule-based algorithms and four heuristic algorithms,
namely, min-min [54], max-min [54], simulated annealing
[55], genetic algorithm [56], particle swarm optimization
[57], and ant colony optimization [58]. For the data sets e-
Economic, e-Protein, and fMRI, the results in Table 2 show
that heuristics usually can find better results than tradi-
tional scheduling algorithms (i.e., min-min, max-min, and
FIFO) in terms of makespan. In addition, the results of ACO
are similar to those of the proposed algorithm for the first
three data sets shown in Table 2. For the large data sets, j30,
j60, and j90, the results of four heuristics (SA, GA, PSO, and
ACO) are better than min-min, max-min, and FIFO. More-
over, the results also show that HHSA outperforms all the
other scheduling algorithms, namely, min-min, max-min,
FIFO, SA, GA, PSO, and ACO for the last three data sets
shown in Table 2.
The results of best-so-far makespan given in Fig. 6 show
that the min-min and max-min get almost exactly the same
makespan in all iterations. For the SA, the higher the tem-
perature which will get a larger fine-tune rate for the current
solution, the poorer the best-so-far makespan compared to
the other non-traditional algorithms at the early iterations.
On the other hand, with a lower temperature, SA will con-
verge quickly. For the population-based heuristic algo-
rithms GA, PSO, and ACO, the search directions will
converge in about 500 iterations. Fig. 6 also shows that
HHSA converges faster than the other heuristic algorithms,
and it usually takes no more than 200 iterations. In addition,
because HHSA is able to automatically choose the suitable
low-level heuristic algorithm and use the perturbation
method to improve the results, the end results of HHSA are
usually better than the other scheduling algorithms, espe-
cially for complex and large-scale problems. A closer look at
these simulation results shows that the max-min algorithm
does not scale well as the number of tasks increases. For
example, for small data sets, the results depicted in Figs. 6a,
6b, and 6c show that max-min can provide a result that is
close to the other heuristic scheduling algorithms in terms
of the makespan (more precisely, the differences are no
more than 3). However, as the number of tasks increases,
the results described in Figs. 6d, 6e, and 6f show that the dif-
ferences between max-min and other heuristic scheduling
algorithms are also increased. For instance, Fig. 6d shows
that the differences between max-min and other heuristic
scheduling algorithms are more than five; Figs. 6e and 6f
show that the differences between them are more than 10.
On the other hand, for the min-min algorithm, the results
given in Figs. 6d, 6e and 6f show that min-min is closer to
(a)
(b)
Fig. 5. Example showing how CloudSim works by assigning tasks to
VMs. (a) virtual machines. (b) workflow tasks.
TABLE 1
Parameter Settings of the Five Cloud Scheduling Algorithms
for Workflow Scheduling on Cloudsim

the other heuristic scheduling algorithm than max-min for
large-scale data sets. Fig. 6 also provides the convergence
information of the scheduling algorithms compared in this
study. The results given in Fig. 6a show that the search pro-
cess of these scheduling algorithms will converge to a stable
state very quickly when the data set is small. A good exam-
ple is the results of ACO and HHSA, which show that the
result at iteration 100 is very close to the result at iteration
1,000. But the results depicted in Fig. 6b show that the
search process of ACO and HHSA can still find a better
result after iteration 100, meaning that the heuristic schedul-
ing algorithms still have a chance to find a better result at
later iterations when the data set becomes large. In Figs. 6c
to 6f, this kind of results becomes quite clear. In addition,
based on the observations of the convergence of these
scheduling algorithms, we can easily know their perfor-
mance. For example, for small data sets, the convergence
speeds of ACO and HHSA described in Figs. 6a, 6b, and 6c
are quite close to each other. However, as the number of
tasks increases, the convergence speeds of ACO and HHSA
are as shown in Figs. 6d, 6e, and 6f. They show that HHSA
has a higher chance to find a better result than ACO does
because it keeps finding a better result after the ACO has
reached a stable state.
5.2 Hadoop Map-Task Scheduling
5.2.1 Data Sets and Parameter Settings
Another distributed system, Hadoop, provides slots as the
resources to be allocated to the independent jobs, as shown
in Fig. 7. Each job will be divided into multiple independent
TABLE 2
Simulation Results of the Workflow Scheduling Algorithms Evaluated in this Study
Fig. 6. Convergence analysis by best-so-far makespan.

map tasks according to a predefined map size, and the num-
ber of reduce tasks is given when submitting the jobs to the
Hadoop cluster. Each scheduler is carried out for 30 runs,
with Hadoop restarted at the end of each run. The maxi-
mum number of iterations each run is set equal to 50; the
population size is set equal to 10 for PBHAs. The parameter
settings of the scheduling algorithms of Hadoop are slightly
different from those used in the simulations on CloudSim
and are as shown in Table 3. By the nature of most evolu-
tionary algorithms, the computation of fitness (makespan in
this case) is required. In order to compute the makespan,
the algorithm needs to know the job size and the processing
speed. For the simulations, both attributes are given in the
data set. But on Hadoop users can only acquire the size of
each map task. The size of each reduce task remains
unknown. Since the makespan of each reduce task cannot
be predicted, only map task scheduling is applicable. There-
fore, for the proposed algorithm, our focus is on scheduling
tasks to map slots.
As mentioned previously, in addition to the size of each
map task, we also need the processing speed of each task.
To obtain this critical attribute, the algorithm first checks if
the job exists in the record of processing speed. If not, one of
the tasks in the job will be prioritized and scheduled imme-
diately. After the map task is finished, we can divide the
size of the map task by the processing time to get the speed
which then can be used to calculate the processing time of
the rest of the map tasks and to predict the total makespan.
In order to simulate the production environment, the job
inter-arrival times in the experiment are roughly exponen-
tially distributed with a mean of 14 seconds, as randomly
sampled from a Facebook trace [41]. The data sets for the
grep jobs on Hadoop are strings each of size 100 bytes and
each on a line by itself randomly generated using a Python
script. The sizes and submission periods are also randomly
generated based on the statistics described in [41]. Fifty jobs
are submitted in the experiment, and the total submission
schedule is 13 minutes long. The sizes and the distributions
of the jobs are as listed in Table 4. In this study, we also tried
submitting larger jobs in a shorter period. More precisely, in
the test, jobs to sort two 1 GB, two 3 GB, two 5 GB, and one
10 GB files are submitted in 1 minute.
In this study, Hadoop was running on both single-node
(machine) and multi-node setups in a virtualized environ-
ment. The single-node setup is an Intel i7-3820 with 64 G
RAM, which runs Gentoo Linux with kernel version 3.7.10.
Hadoop was compiled from source version 1.1.3 using
sun-jdk-1.6 toolchain. By properly configuring the map
and task capacities of Hadoop, a single machine can still
provide multiple working slots. For example, a machine
with eight map task capacities is equivalent to a cluster
consisting of four machines each of which provide two
map task capacities. Meanwhile, each slot can be treated as
an independent virtual machine in the Cloud environment,
which can also be mapped to the virtual machines of
CloudSim in the simulation.
For the multi-node setup, the Hadoop cluster was run-
ning on four virtual machines hosted on a cluster with
AMD Opteron 6128 processors running XenServer. Each
virtual machine has eight cores and 16 GB RAM and runs
CentOS with kernel version 2.6.32. Hadoop was compiled
from source version 1.1.3 using sun-jdk-1.6 toolchain. One
of the four virtual machines is configured as the worker and
job tracker node for scheduling jobs. The other three are
configured as worker nodes. Each node is configured with
four map slots, implying that there are total 16 map slots.
This study uses the map/reduce grep program distrib-
uted with Hadoop for the benchmark. The data sets are ran-
domly generated so that each line contains 100 bytes of
data. To evaluate the performance of HHSA, the same data
sets are fed to FIFO and Fair Scheduler [41]. For larger data
sets, TeraSort [59], which is also provided by Hadoop, is
used to generate and sort the data. Moreover, another data
set, retrieved from the log of 1998 World Cup website [60]
provided by Internet Traffic Archive, for analyzing the web
user access behavior is used to evaluate the performance of
the proposed algorithm when it is applied to a real complex
scheduling problem. In this experiment, the data are used
by a map/reduce web log analyzer to calculate the reference
counts of URLs.
5.2.2 Simulation Results of Hadoop Map-Task
Scheduling
The experimental results of Hadoop are as shown in Table 5
and Table 6. For the grep jobs, the three schedulers give the
similar results in both the single-node and multi-node
setup. The results show that for the grep jobs, the Hadoop
cluster was under-loaded, meaning that when new jobs
arrived, previous jobs and tasks had already finished, thus
making the scheduling results, i.e., the performance, of
three schedulers similar to each other. This is justified by
TABLE 4
Data Set Generated for Benchmark
Fig. 7. Job executing on Hadoop.
TABLE 3
Parameters of the Five Cloud Scheduling Algorithms Evaluated

the average task waiting time depicted in Fig. 10, to be dis-
cussed in later section.
When facing a burst, HHSA can make a better scheduling
decision. For the TeraSort jobs, HHSA outperforms the
other two schedulers. In the single-node setup, HHSA beats
Fair Scheduler and FIFO by 17.19 and 30.75 percent, respec-
tively, on average. In the multi-node setup, HHSA again
outperforms the other two schedulers. On average, HHSA
is 11.36 percent faster than FIFO and 4.56 percent faster
than Fair Scheduler. As the results show, by optimizing the
makespan, HHSA can provide a better scheduling decision
in terms of the best-so-far makespan. The results of web log
analyzer show that HHSA performs marginally better than
FIFO, by 2:62 and 0:25 percent, respectively, in the single-
node and multi-node setup. This is due to the fact that the
web log analyzing job takes OðnÞ time; thus, the time costs
for all the tasks are very close to each other. Under this cir-
cumstance, FIFO can easily produce a good schedule that is
close to that produced by the proposed algorithm. On the
other hand, the time cost of each TeraSort task varies
greatly, thus largely increasing the difficulty of scheduling.
In this case, HHSA outperforms the other two schedulers
significantly.
The performance of FIFO and HHSA on CloudSim and
Hadoop is as shown in Table 7. For both CloudSim and
Hadoop, a set of jobs the size of which are close to those of
TeraSort benchmarks are generated and used. Then, the
size of each task is randomly generated so that it falls in the
range of 64 MB Æ 30%; each job will be divided into a set of
smaller tasks of the same size though the processing time of
each task is still not necessarily the same. When most tasks
are of the same size, FIFO, which uses greedy strategy to
scheduling tasks, can achieve a high utilization and good
makespan. However, when the tasks are of different size,
FIFO will become the worst method. Moreover, the simula-
tion results also show that HHSA outperforms FIFO, espe-
cially when the problem is complex. The results are
consistent with those of Table 2, where the sizes of jobs in
the data sets for CloudSim are more diverse than those in
this experiment.
5.3 Analysis
The maximum number of iterations of LLH fmax and the
maximum number of non-improved iterations fni set up for
the workflow scheduling problem may not be suitable for
the Hadoop map-task scheduling problem. As such, to eval-
uate the performance of HHSA for the Hadoop map-task
scheduling problem, four data sets with a varying number
of independent tasks, namely, 25 tasks, 50 tasks, 100 tasks,
and 200 tasks, are used. Here, our focus is on the average
makespan of 30 runs with fmax in the range of 20 to C and
fni in the range of 1 to C where C is the maximum number
of iterations of HHSA. This experiment is aimed at measur-
ing the performance of HHSA when applying to the
Hadoop map-task scheduling problem with the population
size fixed at 50 for the PBHAs. This eventually takes into
consideration all the combinations of the parameter settings
shown in Table 8.
5.3.1 The Maximum Number of Iterations of LLH
Fig. 8 shows the makespans for a varying maximum num-
ber of iterations of LLH fmax in the range ½20; CŠ and a vary-
ing maximum number of iterations C 2 f250; 500; 750;
1;000g for each data set. It is worth noting that the maxi-
mum number of non-improved iterations fni is set equal to
fmax. This test is aimed to help us understand whether the
value of fmax will influence the scheduling result. As the
results in Fig. 8 show, the larger the number of iterations,
the better the makespan. For instance, 250 iterations give a
worse result than the other three cases. The results further
show that the lower the maximum number of iterations of
low-level algorithm fmax, the better the makespan. Appar-
ently, as the value of fmax approaches the maximum num-
ber of iterations, it implies that the probability of using LLH
is getting smaller. In the worse case, only one of the LLH’s
will be used. That is, HHSA will degenerate to one of the
low-level heuristics, which in turn will reduce the chance of
balancing the intensification and diversification in the
TABLE 5
Performance of HHSA on Hadoop with Single-Node
(in Seconds)
TABLE 6
Performance of HHSA on Hadoop with Multi-Node (in Seconds)
TABLE 7
Performance of HHSA on CloudSim and Hadoop (in Seconds)
TABLE 8
Parameter Settings for the Analysis

search of solutions. In addition, Figs. 8b, 8c and 8d show
that for problems larger than that given in Fig. 8a, a similar
result is obtained; that is, decreasing the value of fmax for
the proposed algorithm can provide better results in most
cases in terms of makespan. Consequently, as Fig. 8 shows,
the smaller the value of fmax, the better the makespan.
5.3.2 The Maximum Number of Non-Improved
Iterations
Furthermore, Fig. 9 shows the makespans for a varying maxi-
mum number of non-improved iterations fni 2 f1; 2; . . . ; Cg
and a varying maximum number of iterations C 2 f250;
500; 750; 1;000g for each data set. In order to eliminate the
inﬂuence the maximum number of iterations of the selected
LLH may have, the value of fmax is always set equal to the
maximum number of iterations C. When the maximum num-
ber of non-improved iterations is set equal to a small value, it
means that if the best-so-far makespan of the selected LLH is
not improved within the given number of iterations, the
selected LLH will be replaced by another LLH. This implies
that a smaller value of fni will increase the diversity of the
search directions of the selected LLH, thus improving the
quality for not only the small problem shown in Fig. 9a but
also the larger and complex problems depicted in Figs. 9b, 9c,
Fig. 8. Average makespan for a varying number of tasks and fmax.
Fig. 9. Average makespans for a varying number of tasks and fni.

and 9d. In contrast, a larger value of fni will increase the
speed of convergence. Since the diversity and the quality of
makespan may not be improved at the same time, our obser-
vation shows that a suitable value of fni can not only get a bet-
ter mean makespan, it also uses LLH more efficiently.
5.3.3 The Average Task Waiting Time
The average task waiting times for grep and TeraSort in the
single-node and multi-node setup on Hadoop are as shown
in Fig. 10. The small waiting times of grep given in Fig. 10a
imply that the cluster is under-loaded and thus most tasks
can be scheduled immediately. These results also explain
that for small data sets in a cluster that is under-loaded, the
performance of these scheduling algorithms is pretty much
the same. On the other hand, the results of TeraSort
described in Fig. 10b indicate that for large data sets in a
cluster that is fully loaded, compared to FIFO, HHSA can
reduce the average task waiting time of TeraSort by 29:042
and 12:49 percent,3
respectively, in the single-node and
multi-node setup. This means that by optimizing the make-
span, tasks in the queue can be completed earlier; thus,
incoming tasks can be scheduled sooner and the average
task waiting time can be further reduced. Moreover, the
results of HHSA for TeraSort on Hadoop show that HHSA
takes only 0.0144 milliseconds on average, meaning that the
computation time of HHSA is negligible in most cases.
5.3.4 The Cloud Rental Payment
In terms of makespan, on Hadoop, HHSA is about 11.36
percent faster than FIFO on average. The multi-node setup
used in this paper is based on the “Standard On-Demand
Extra Large Instances” of Amazon Elastic MapReduce ser-
vice [61], which costs 0:12 US dollars per hour. Though the
proposed algorithm described herein is aimed at optimizing
the makespan instead of the budget. The reality is that by
reducing the makespan, it also reduces the total cost of
cloud service. According to the pricing and taking into
account the improvement of the makespan, the proposed
method can reduce 0:12 Â 11:36% ¼ 0:014 US dollars
per hour for each standard on-demand extra large instance.
Although HHSA takes a longer computation time than the
other schedulers, the difference is negligible because it is
compensated by a better makespan, which will save much
more time than HHSA takes.
5.4 Summary
As mentioned in Section 2.1.2, the characteristics of a cloud
system are quite different from those of the information sys-
tems available today. A typical characteristic of a cloud sys-
tem is in that the number of tasks it has to handle is much
larger than that an information system available today has
to handle, thus making it hard to schedule all the tasks.
Another typical characteristic that affects the design of
scheduling algorithms for a cloud system is in that a cloud
system usually contains a number of virtual machines (or
nodes) that add more uncertainties for task scheduling. Still
another characteristic is in that a cloud system usually has
many heterogeneous tasks to be scheduled, thus making it
much more difficult to schedule than tasks on the informa-
tion systems available today.
To solve these problems, a high-performance scheduling
algorithm which can provide a better scheduling result in a
reasonable time is needed. This explains why a hyper-
heuristic-based scheduling algorithm is presented in this
paper, which uses various search strategies to solve the
scheduling problem of a cloud for providing a better result
than traditional scheduling algorithms and other heuristic
scheduling algorithms do. Because the design of the pro-
posed algorithm is not for a particular cloud system, it can
be applied to many applications that require task schedul-
ing, such as science computing on Cloud. But the high
latency inherent in the heuristic-based algorithms will make
them unsuitable for the low-level scheduling of IaaS. More-
over, to the best of our knowledge, there exists no general
and efficient way to automatically determine the number of
resources that is used for each application for a very simple
reason. The number of resources required may differ for dif-
ferent input to the application.
Compared to most traditional rule-based deterministic
scheduling algorithms (e.g., min-min, max-min, and FIFO)
that use only one search direction, the proposed algorithm
uses multiple search directions at each iteration during the
convergence process, thus having a higher chance to find a
better result than the rule-based deterministic algorithms
do. The results depicted in Sections 5.1.2 and 5.2.2 and
Table 6 show that in terms of makespan, the proposed algo-
rithm outperforms the traditional scheduling algorithms in
most cases, for not only the single-node but also the multi-
node data sets. Compared to the other heuristic scheduling
algorithms, the proposed algorithm, again in terms of make-
span, can still provide a better result than the other heuristic
scheduling algorithms, in terms of not only the results of
CloudSim but also the results of Hadoop, as shown in
Sections 5.1.2 and 5.2.2. One of the reasons that can be easily
justified is that HHSA leverages the strengths of many
search strategies provided by different heuristic scheduling
algorithms (low-level algorithms), thus making it more
diverse in terms of the directions for which it searches dur-
ing the convergence process, especially when compared to
the other heuristic scheduling algorithms.
Fig. 10. Average task waiting times for grep and TeraSort.
2. ½ð374:90 À 266:03Þ=374:90Š Â 100 ¼ 29:04%.
3. ½ð432:60 À 378:56Þ=432:60Š Â 100 ¼ 12:49%.

For the stability issue, HHSA is not designed for inter-
connectivity and locality because they will degrade the per-
formance of Hadoop. As the authors of [62] pointed out, if
the nodes of a virtualized Hadoop cluster actually reside in
the same physical node, the overhead incurred for transfer-
ring data between these nodes is generally smaller than
transferring the data between physical nodes. Hence, if the
overhead of connection is too large but the scheduling prob-
lem is too small, the performance of the proposed algorithm
will be degraded; otherwise, the performance of the pro-
posed algorithm will be better than the other scheduling
algorithms, especially when the scheduling problem
becomes too complex or too large.
6 CONCLUSIONS
This study presents a high-performance hyper-heuristic
algorithm to find better scheduling solutions for cloud com-
puting systems. The proposed algorithm uses two detection
operators to automatically determine when to change the
low-level heuristic algorithm and a perturbation operator to
fine-tune the solutions obtained by each low-level algorithm
to further improve the scheduling results in terms of make-
span. As the simulation results show, the proposed algo-
rithm can not only provide better results than the
traditional rule-based scheduling algorithms, it also outper-
forms the other heuristic scheduling algorithms, in solving
the workflow scheduling and Hadoop map-task scheduling
problems on cloud computing environments. In addition,
the simulation results show further that the proposed algo-
rithm converges faster than the other heuristic algorithms
evaluated in this study for most of the data sets. In brief, the
basic idea of the proposed “hyper-heuristic” algorithm is to
leverage the strengths of all the low-level algorithms while
not increasing the computation time, by running one and
only one low-level algorithm at each iteration. This is funda-
mentally different from the so-called hybrid heuristic algo-
rithm, which runs more than one low-level algorithm at
each iteration, thus requiring a much longer computation
time. Our experiences show that the proposed algorithm
can be applied to different cloud computing systems to
enhance the performance of scheduling problems. In the
future, our focus will be on finding more effective detection
operators and perturbation method to enhance the perfor-
mance of the proposed algorithm. We will also attempt to
apply HHSA to different research domains.
ACKNOWLEDGMENTS
The authors would like to thank the editors and anonymous
reviewers for their valuable comments and suggestions on
the paper that greatly improve the quality of the paper. This
work was supported in part by the National Science Council
of Taiwan, R.O.C., under Contracts NSC102-2221-E-041-006,
NSC102-2221-E-110-054, and NSC102-2219-E-006-001.
REFERENCES
[1] M. Pinedo, Scheduling: Theory, Algorithms and Systems. Englewood
Cliffs, NJ, USA: Prentice Hall, 1995.
[2] Scheduling Theory and Its Applications, P. Chretienne, E. G. Coff-
man, J. K. Lenstra, and Z. Liu, eds., Hoboken, NJ, USA: Wiley,
1995.
[3] A. Allahverdi, C. T. Ng, T. C. E. Cheng, and M. Y. Kovalyov, “A
survey of scheduling problems with setup times or costs,” Eur. J.
Oper. Res., vol. 187, no. 3, pp. 985–1032, 2008.
[4] M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A. Konwin-
ski, G. Lee, D. Patterson, A. Rabkin, I. Stoica, and M. Zaharia, “A
view of cloud computing,” Commun. ACM, vol. 53, no. 4, pp. 50–58,
2010.
[5] I. Foster, Y. Zhao, I. Raicu, and S. Lu, “Cloud computing and grid
computing 360-degree compared,” in Proc. Grid Comput. Environ.
Workshop, 2008, pp. 1–10.
[6] K. Bilal, M. Manzano, S. Khan, E. Calle, K. Li, and A. Zomaya, “On
the characterization of the structural robustness of data center
networks,” IEEE Trans. Cloud Comput., vol. 1, no. 1, pp. 64–77, Jan.–
Jun. 2013.
[7] X. Lei, X. Liao, T. Huang, H. Li, and C. Hu, “Outsourcing large
matrix inversion computation to a public cloud,” IEEE Trans.
Cloud Comput., vol. 1, no. 1, pp. 78–87, Jan. 2013.
[8] M. R. Garey and D. S. Johnson, Computer and Intractability: A Guide
to the Theory of NP-Completeness. San Francisco, CA, USA: Free-
man, 1979.
[9] C. W. Tsai and J. Rodrigues, “Metaheuristic scheduling for cloud:
A survey,” IEEE Syst. J., vol. 8, no. 1, pp. 279–297, Mar. 2014.
[10] S. Kirkpatrick, C. D. Gelatt, and M. P. Vecchi, “Optimization by
simulated annealing,” Science, vol. 220, no. 4598, pp. 671–680, 1983.
[11] J. H. Holland, Adaptation in Natural and Artificial Systems: An Intro-
ductory Analysis with Applications to Biology, Control, and Artificial
Intelligence. Ann Arbor, MI, USA: Univ. of Michigan Press, 1975.
[12] J. Kennedy and R. C. Eberhart, “Particle swarm optimization,” in
Proc. IEEE Int. Conf. Neural Netw., 1995, pp. 1942–1948.
[13] M. Dorigo and L. M. Gambardella, “Ant colony system: A cooper-
ative learning approach to the traveling salesman problem,” IEEE
Trans. Evol. Comput., vol. 1, no. 1, pp. 53–66, Apr. 1997.
[14] R. N. Calheiros, R. Ranjan, A. Beloglazov, C. A. De Rose, and R.
Buyya, “CloudSim: A toolkit for modeling and simulation of
cloud computing environments and evaluation of resource provi-
sioning algorithms,” Softw.: Practice Experience, vol. 41, no. 1,
pp. 23–50, 2011.
[15] J. B»a_zewicz, K. H. Ecker, E. Pesch, G. Schmidt, and J. Wȩglarz,
Scheduling Computer and Manufacturing Processes. New York, NY,
USA: Springer, 2001.
[16] Y. T. J. Leung, Handbook of Scheduling: Algorithms, Models and Per-
formance Analysis. London, U.K.: Chapman Hall, 2004.
[17] K. M. Elsayed and A. K. Khattab, “Channel-aware earliest dead-
line due fair scheduling for wireless multimedia networks,” Wire-
less Personal Commun., vol. 38, no. 2, pp. 233–252, 2006.
[18] W. H. Kohler, “A preliminary evaluation of the critical path
method for scheduling tasks on multiprocessor systems,” IEEE
Trans. Comput., vol. C-24, no. 12, pp. 1235–1238, Dec. 1975.
[19] R. Ingalls and D. Morrice, “PERT scheduling with resources using
qualitative simulation graphs,” in Proc. Simul. Conf., 2000, vol. 1,
pp. 362–370.
[20] H.-J. Sch€utz and R. Kolisch, “Approximate dynamic program-
ming for capacity allocation in the service industry,” Eur. J. Oper.
Res., vol. 218, no. 1, pp. 239–250, 2012.
[21] D. Shi and T. Chen, “Optimal periodic scheduling of sensor net-
works: A branch and bound approach,” Syst. Control Lett., vol. 62,
no. 9, pp. 732–738, 2013.
[22] S. M. Johnson, “Optimal two- and three-stage production sched-
ules with setup times included,” Naval Res. Logistics Quart., vol. 1,
no. 1, pp. 61–68, 1954.
[23] M. R. Garey, D. S. Johnson, and R. Sethi, “The complexity of flow-
shop and jobshop scheduling,” Math. Oper. Res., vol. 1, no. 2,
pp. 117–129, 1976.
[24] D. Daniel and S. J. Lovesum, “A novel approach for scheduling
service request in cloud with trust monitor,” in Proc. Int. Conf. Sig-
nal Process., Commun., Comput. Netw. Technol., 2011, pp. 509–513.
[25] Z. Wu, X. Liu, Z. Ni, D. Yuan, and Y. Yang, “A market-oriented
hierarchical scheduling strategy in cloud workflow systems,” The
J. Supercomput., vol. 63, no. 1, pp. 256–293, 2013.
[26] B. Saovapakhiran, G. Michailidis, and M. Devetsikiotis,
“Aggregated-DAG scheduling for job flow maximization in het-
erogeneous cloud computing,” in Proc. IEEE Global Telecommun.
Conf., 2011, pp. 1–6.
[27] M. Rahman, X. Li, and H. Palit, “Hybrid heuristic for scheduling
data analytics workflow applications in hybrid cloud environ-
ment,” in Proc. IEEE Int. Symp. Parallel Distrib. Process. Workshops,
2011, pp. 966–974.

[28] Apache Hadoop. [Online]. Available: http://hadoop.apache.org,
2014.
[29] Y. M. Huang and Z. H. Nie. Cloud computing with Linux and
Apache Hadoop. [Online]. Available: http://www.ibm.com/
developerworks/aix/library/au-cloud_apache/, 2014.
[30] B. David. Open source cloud computing with Hadoop. [Online].
Available: http://www.linuxjournal.com/content/open-source-
cloud-computing-hadoop, 2014.
[31] A. O’Driscoll, J. Daugelaite, and R. D. Sleator, “‘Big data’, Hadoop
and cloud computing in genomics,” J. Biomed. Inf., vol. 46, no. 5,
pp. 774–781, 2013.
[32] Y. Song, M. Wu, and L. Yang, “Study of smart grid marketing sys-
tem architecture based on Hadoop platform of cloud computing,”
J. Commun. Comput., vol. 9, pp. 741–743, 2012.
[33] A. Bala and I. Chana, “A survey of various workflow scheduling
algorithms in cloud environment,” in Proc. Nat. Conf. Inf. Commun.
Technol., 2011, pp. 26–30.
[34] N. Kaur, T. S. Aulakh, and R. S. Cheema, “Comparison of work-
flow scheduling algorithms in cloud computing,” Int. J. Adv. Com-
put. Sci. Appl., vol. 2, no. 10, pp. 81–86, 2011.
[35] J. Li, M. Qiu, Z. Ming, G. Quan, X. Qin, and Z. Gu, “Online optimi-
zation for scheduling preemptable tasks on IaaS cloud systems,” J.
Parallel Distrib. Comput., vol. 72, no. 5, pp. 666–677, 2012.
[36] J. Li, M. Qiu, J. Niu, Y. Chen, and Z. Ming, “Adaptive resource
allocation for preemptable jobs in cloud systems,” in Proc. Int.
Conf. Intel. Syst. Design Appl., 2010, pp. 31–36.
[37] C. Lin and S. Lu, “Scheduling scientific workflows elastically for
cloud computing,” in Proc. IEEE Int. Conf. Cloud Comput., 2011,
pp. 746–747.
[38] S. Abrishami and M. Naghibzadeh, “Deadline-constrained work-
flow scheduling in software as a service cloud,” Scientia Iranica,
vol. 19, no. 3, pp. 680–689, 2012.
[39] Y. C. Lee, C. Wang, A. Y. Zomaya, and B. B. Zhou, “Profit-driven
scheduling for cloud services with data access awareness,” J. Par-
allel Distrib. Comput., vol. 72, no. 4, pp. 591–602, 2012.
[40] K. Kousalya and P. Balasubramanie, “Ant algorithm for grid
scheduling powered by local search,” Int. J. Open Problem Comput.
Sci. Math., vol. 1, no. 3, pp. 222–240, 2008.
[41] M. Zaharia, D. Borthakur, J. Sen Sarma, K. Elmeleegy, S. Shenker,
and I. Stoica, “Delay scheduling: A simple technique for achieving
locality and fairness in cluster scheduling,” in Proc. Eur. Conf.
Comput. Syst., 2010, pp. 265–278.
[42] Yahoo!. Hadoop capacity scheduler. [Online]. Available:
http://hadoop.apache.org/docs/r2.3.0/hadoop-yarn/hadoop-
yarn-site/CapacityScheduler.html, 2014.
[43] H. H. Hoos and T. St€utzle, Stochastic Local Search: Foundations
Applications. San Mateo, CA, USA: Morgan Kaufmann, 2004.
[44] P. Cowling and G. Kendall, “A hyperheuristic approach to sched-
uling a sales summit,” in Proc. Practice Theory Autom. Timetabling
III, 2001, pp. 176–190.
[45] P. I. Cowling, G. Kendall, and E. Soubeiga, “A hyperheuristic
approach to scheduling a sales summit,” in Proc. Practice Theory
Autom. Timetabling, 2000, pp. 176–190.
[46] B. K. Ambati, J. Ambati, and M. M. Mokhtar, “Heuristic combina-
torial optimization by simulated Darwinian evolution: A polyno-
mial time algorithm for the traveling salesman problem,” Biol.
Cybern., vol. 65, no. 1, pp. 31–35, 1991.
[47] L. Y. Tseng and S. B. Yang, “A genetic approach to the automatic
clustering problem,” Pattern Recognit., vol. 34, no. 2, pp. 415–424,
2001.
[48] D. Laha and U. Chakraborty, “An efficient hybrid heuristic for
makespan minimization in permutation flow shop scheduling,”
The Int. J. Adv. Manuf. Technol., vol. 44, no. 5/6, pp. 559–569, 2009.
[49] L. Ramakrishnan, C. Koelbel, Y.-S. Kee, R. Wolski, D. Nurmi, D.
Gannon, G. Obertelli, A. YarKhan, A. Mandal, T. M. Huang, K.
Thyagaraja, and D. Zagorodnov, “VGrADS: Enabling e-science
workflows on grids and clouds with fault tolerance,” in Proc.
Conf. High Performance Comput. Netw., Storage Anal., 2009,
pp. 47:1–47:12.
[50] M. Mao and M. Humphrey, “Auto-scaling to minimize cost
and meet application deadlines in cloud workflows,” in Proc.
Int. Conf. High Performance Comput., Netw., Storage Anal., 2011,
pp. 1–12.
[51] Y. Zhao, M. Wilde, I. Foster, J. Voeckler, T. Jordan, E. Quigg, and J.
Dobson, “Grid middleware services for virtual data discovery,
composition, and integration,” in Proc. Workshop Middleware Grid
Comput., 2004, pp. 57–62.
[52] A. O’Brien, S. Newhouse, and J. Darlington, “Mapping of scien-
tific workflow within the e-protein project to distributed
resources,” in Proc. UK e-Science All Hands Meeting, Nottingham,
U.K., 2004, pp. 404–409.
[53] R. Kolisch and A. Sprecher, “PSPLIB—A project scheduling problem
library: OR software—ORSEP operations research software
exchange program,” Eur. J. Oper. Res., vol. 96, no. 1, pp. 205–216,
1997.
[54] J. Blythe, S. Jain, E. Deelman, Y. Gil, and K. Vahi, “Task schedul-
ing strategies for workflow-based applications in grids,” in Proc.
IEEE Int. Symp. Cluster Comput. Grid, 2005, pp. 759–767.
[55] S. Benedict and V. Vasudevan, “Scheduling of scientific work-
flows using simulated annealing algorithm for computational
grids,” Int. J. Soft. Comput., vol. 2, no. 5, pp. 606–611, 2007.
[56] J. Yu and R. Buyya, “A budget constrained scheduling of work-
flow applications on utility grids using genetic algorithms,” in
Proc. Workflows Support Large-Scale Sci., 2006, pp. 1–10.
[57] T. Chen, B. Zhang, X. Hao, and Y. Dai, “Task scheduling in grid
based on particle swarm optimization,” in Proc. Int. Symp. Parallel
Distrib. Comput., 2006, pp. 238–245.
[58] W. N. Chen and J. Zhang, “An ant colony optimization approach
to a grid workflow scheduling problem with various QoS
requirements,” IEEE Trans. Syst., Man, Cybern., Part C: Appl. Rev.,
vol. 39, no. 1, pp. 29–43, Jan. 2009.
[59] O. O’Malley. TeraByte Sort on Apache Hadoop. [Online]. Avail-
able: http://sortbenchmark.org/YahooHadoop.pdf, 2014.
[60] I. T. Archive. World Cup 98. [Online]. Available: http://ita.ee.lbl.
gov/html/contrib/WorldCup.html, 2014.
[61] Amazon, Amazon elastic mapreduce. [Online]. Available: http://
aws.amazon.com/elasticmapreduce, 2014.
[62] H. S. Shin, K. H. Kim, C. Y. Kim, and S. I. Jung, “The new approach
for inter-communication between guest domains on virtual machine
monitor,” in Proc. Int. Symp. Comput. Inf. Sci., 2007, pp. 1–6.
Chun-Wei Tsai received the MS degree in
management information system from National
Pingtung University of Science and Technology,
Pingtung, Taiwan, in 2002, and the PhD degree
in computer science from National Sun Yat-sen
University, Kaohsiung, Taiwan, in 2009. He was
also a postdoctoral fellow with the Department of
Electrical Engineering, National Cheng Kung
University, Tainan, Taiwan before joining the
faculty of Applied Geoinformatics, and then the
faculty of Information Technology, Chia Nan Uni-
versity of Pharmacy Science, Tainan, Taiwan, in 2010 and 2012,
respectively, where he is currently an assistant professor. His research
interests include information retrieval, evolutionary computation, internet
technology, and combinatorial optimization.
Wei-Cheng Huang received the BS degree in
computer science and information engineering
from the National Pingtung Institute of Com-
merce, Pingtung, Taiwan, in 2010, and the MS
degree in computer science and engineering
from National Sun Yat-sen University, Kaoh-
siung, Taiwan, in 2012. He is currently working
toward the PhD degree in computer science
and engineering from National Sun Yat-sen
University. His current research interests
include evolutionary computation, clustering,
scheduling, and cloud computing.

Meng-Hsiu Chiang received the BS degree in
computer science and engineering from National
Sun Yat-sen University, Kaohsiung, Taiwan, in
2010. He is currently working toward the MS
degree in computer science and engineering
from National Sun Yat-sen University. His current
research interests include scheduling, distributed
systems, and cloud computing.
Ming-Chao Chiang received the BS degree in
management science from National Chiao
Tung University, Hsinchu, Taiwan, ROC, in
1978, and the MS, MPhil, and PhD degrees in
computer science from Columbia University,
New York, NY, in 1991, 1998, and 1998,
respectively. He has more than 12 years of
experience in the software industry encom-
passing a wide variety of roles and responsi-
bilities in both large and start-up companies
before joining the faculty of the Department of
Computer Science and Engineering, National Sun Yat-sen Univer-
sity, Kaohsiung, Taiwan, ROC, in 2003, where he is currently an
associate professor. His current research interests include image
processing, evolutionary computation, and system software.
Chu-Sing Yang received the BS degree in engi-
neering science, and the MS and PhD degrees in
electrical engineering from National Cheng Kung
University, Tainan, Taiwan, in 1976, 1984 and
1987, respectively. He joined the faculty of the
Department of Electrical Engineering, National
Sun Yat-sen University, Kaohsiung, Taiwan, as
an associate professor, in 1988. Since 1993, he
has been a professor of the Department of Com-
puter Science and Engineering, National Sun
Yat-sen University. He was a chair of Department
of Computer Science and Engineering, National Sun Yat-sen University
from August 1995 to July 1999, a director of Computer Center, National
Sun Yat-sen University from August 1998 to October 2002, and a pro-
gram chair of ICS-96 and Program Co-Chair of ICPP-2003. He joined
the faculty of the Department of Electrical Engineering, National Cheng
Kung University, Tainan, Taiwan, as a professor in 2006. His research
interests include web mining, embedded system, network management,
and wireless communication.
For more information on this or any other computing topic,
please visit our Digital Library at www.computer.org/publications/dlib.

Ieeepro techno solutions 2014 ieee dotnet project - a hyper-heuristic scheduling

More Related Content

What's hot (17)

Viewers also liked (19)

Similar to Ieeepro techno solutions 2014 ieee dotnet project - a hyper-heuristic scheduling (20)

More from ASAITHAMBIRAJAA (13)

Recently uploaded (20)

Ieeepro techno solutions 2014 ieee dotnet project - a hyper-heuristic scheduling