Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review

International Journal in Foundations of Computer Science & Technology (IJFCST) Vol.6, No.1, January 2016
DOI:10.5121/ijfcst.2016.6106 65
CURRENT PERSPECTIVE IN TASK SCHEDULING
TECHNIQUES IN CLOUD COMPUTING: A REVIEW
Jyoti Thaman, Manpreet Singh
Department of Computer Science & Engineering M. M. University, Ambala,
Haryana, India
ABSTRACT
Cloud computing is a development of parallel, distributed and grid computing which provides computing
potential as a service to clients rather than a product. Clients can access software resources, valuable
information and hardware devices as a subscribed and monitored service over a network through cloud
computing.Due to large number of requests for access to resources and service level agreements between
cloud service providers and clients, few burning issues in cloud environment like QoS, Power, Privacy and
Security, VM Migration, Resource Allocation and Scheduling need attention of research
community.Resource allocation among multiple clients has to be ensured as per service level agreements.
Several techniques have been invented and tested by research community for generation of optimal
schedules in cloud computing. A few promising approaches like Metaheuristics, Greedy, Heuristic
technique and Genetic are applied for task scheduling in several parallel and distributed systems. This
paper presents a review on scheduling proposals in cloud environment.
KEYWORDS
Metaheuristics, Heuristics, Greedy, Genetic Algorithms, Review, Task Scheduling.
1. INTRODUCTION
The cloud computing model is composed of three service models and four deployment models as
shown in Figure 1. Clouds offer services that can be grouped into three categories: Software as a
Service (SaaS), Platform as a Service (PaaS) and Infrastructure as a Service (IaaS). Deployment
Models are classified as Public Clouds, Private Clouds, Community Clouds and Hybrid Clouds.
This huge conglomeration of services and sources, shared among clients on subscription basis
needs a serious attention in terms of tasks scheduling, resource allocation and resource sharing.
Anotherimportant concernsfor cloud service providers are energy efficiency and bandwidth
management. If seen by outsider’s view, a cloud environment processes the tasks submitted by
clients. Any simultaneous access to resources needs to be addressed with objectives of improved
resource utilization, reduced energy expenses and reduced makespan. Several proposals in recent
times addressed scheduling issues in cloud environments.
In literature several promising and successful techniques and approaches have evolved as master
tool for guaranteed optimal scheduling of tasks. Scheduling proposals in recent times can be
classified as Metaheuristics, Heuristics, Greedy and Genetic Approach based solutions.
Metaheuristics approach like Particle Swarm Optimization (PSO), Ant Colony Optimization

66
(ACO), BAT echolocation and Artificial Bee Colony (ABC) hasfound their places in scheduling
of tasks in cloud environments. Several approaches like Simulated Annealing, Hill climbing etc.
are classified as heuristic techniques. Heuristics techniques may not guarantee optimal solution
but may provide incidental solution of some problems. Another class of algorithms used genetic
operators for generation and searching local and global optimal solution over solution space. This
category of algorithms may be classified as Genetic Algorithms (GAs). Greedy based solution to
scheduling problem in cloud environment is another valuable approach. This paper has presents a
review of many proposals in recent times and presents a novel taxonomy for classification of
algorithms into Goal Oriented Task Scheduling or Constraints Oriented Task Scheduling.
Figure 1 Cloud Platform
This paper is organized as follows. Section 2 presents a new taxonomy for classification of
scheduling approaches. Section 3 presents a detailed review of scheduling schemes. Finally,
Section 4 concludes the article.
2. TASK SCHEDULING TAXONOMY
Task scheduling is prime concern when a group of clients with specified Service Level
Agreement (SLA) submits their jobs for processing to cloud over network. Tasks are submitted
with certain constraints and deadline requirements. Specified constraints must be respected as per
SLA. Any breach in SLA costs cloud service providers in terms of both economically and
trustworthiness wise. To maintain high trustworthiness and reliability in service, Tasks must be
scheduled with either goal or constraints in mind. When client doesn’t specify any constraint,
cloud service provider must strive to reduce the energy expenses and improve upon resource
utilization. Such scheduling may be considered under Goal Oriented Task Scheduling (GOTS).
Software as a Service (SaaS)
Business Application, Web services & Multimedia
Google
Apps
Face
book
Microsoft cloud
services
Platform as a Service (PaaS)
Software Framework storage (DB/File)
Google
App Engine
Microsoft
Aware
Force.com
Infrastructure as a Service (IaaS)
Computation (VM) storage (block)
Amazon
EC2
IBM VMWare
vSphere
Service
Model
Deployment
Model
Public
Cloud
Private
Cloud
Community
Cloud
Hybrid Cloud

67
GOTS allows the cloud service providers to optimize the resources, by maximizing the resource
usage, minimizing the delays and maximizing the bandwidth usage etc. GOTS allows cloud
service providers to optimize their services and ensure maximum Return on Investment (ROI).
Figure 2 Task Scheduling Taxonomy
In case client specifies certain constraints like cost, deadlines and delays etc. cloud service
providers must schedule tasks of clients with constraints in mind. This type of task scheduling is
called Constraint Oriented Task Scheduling (COTS). COTS give a biased chance to cloud service
providers to negotiate SLA with users on their own terms. Cloud service providers may strike a
better deal in this scenario. The clients who submit their tasks with constraints are prioritized over
other clients and treated as priority costumers. In brief, task scheduling schemes may be classified
either as GOTS orCOTS. Figure 2 presents a classification of task scheduling and present
taxonomy for classification of task scheduling schemes in cloud environment.
Several classical and recent promising scheduling approaches are classified either as COTS or
GOTS or both. Metaheuristics based solutions, heuristic based solutions, genetic algorithm based
solutions and greedy approach based solutions are generally used to implement scheduling in
cloud and grid computing environments. Depending upon the goals achieved or constraints
honoured, this paper reclassified various recent solutions either as COTS or GOTS. This paper
has reviewed various schemes under Metaheuristic, heuristic and genetic algorithm based
approaches.
3. METAHEURISTIC BASED SCHEDULING
Metaheuristic based solutions are proposed using several distinct forms of metaheuristics. Particle
Swarm Optimization (PSO), Ant Colony Optimization (ACO), Artificial Bee Colony (ABC) and
BAT’s echolocation behavior (BAT) are most promising metaheuristic approaches used in task
scheduling.
Task Scheduling
Goal Based
Constrained Based
Metaheuristic Genetic
ACO
Heuristic
PSO
Greedy
ABC BAT

68
3.1 Particle Swarm Optimization (PSO)
PSO has been found to be robust and is successfully applied in solving nonlinear, non-
differentiable multi-modal problems quickly. It is still in its infancy. Many research works have
mentioned application of PSO in task scheduling. PSO is most successful metaheuristic for
generations of optimal scheduling solutions. PSO scans over solution space during each iteration
and accumulates global best and local best solutions. This section presents review of recent
proposals which considered PSO in the field of task scheduling in cloud environment.
Originally PSO was proposed in [1] where PSO was proposed as an optimization tool. Two types
of PSO namely, Discrete PSO and Continuous PSO versions were proposed. With several passes
over the search space and updating local best and global best solutions during each pass, PSO
performed much faster than ACO or GA. In [2] authors introduced the concept of inertia weight
into the original PSO. With introduction of inertia weight PSO could converge even faster.
Initially inertia weight was proposed to lie in the range [0.9, 1.2], which can improve performance
of PSO. Different values of inertia allowed better control over solution search space. Higher
values of inertia weight will result in overshooting the and lower values will trap search in
definite area in search space. A Cost Aware Modified PSO (CA-PSO) was proposed in [3]. In [4]
authors exploit PSO for optimizing overall tasks completion cost in a workflow and respecting the
given deadline constraints. The proposed metaheuristic approach based on PSO succeeds whereas
IC-PCP fails to meet application’s deadline. In comparison IC-PCP failedto meet deadline
constraints as IC-PCP ignored VM boot time. Results prove that PSO performs better than current
state-of-the-art algorithms. Proposal considered deadline constraint. Proposal generates constraint
makespan and performs cost evaluation for various workflows like Montage, Ligo etc. When
compared to SCS, proposed algorithm is capable of generating better schedules and achieved cost
optimization. In [5] authors proposed mathematical model using a Load Balancing Mutation
(balancing) Particle Swarm Optimization (LBMPSO) and considered reliability and availability
as the objective parameters of proposals. LBMPSO used an algorithm to generate schedule and
allocation for cloud computing environment. Algorithm considered available resources for
generation of schedule and allocation patterns. Basic PSO suffers from free VMs, allocation of
more than one task to same VM, allocation of same tasks to multiple VMs and premature
convergence. LBMPSO takes into account execution time, transmission time, make span, round
trip time, transmission cost and load balancing between tasks and achieved reliability in task
scheduling. Idea of LBMPSO is to reschedule failure tasks to available VM. LBMPSO
performance was compared with standard PSO, random algorithm and Longest Cloudlet to
Fastest Processor (LCFP) algorithm to show that LBMPSO can save in make span, execution
time, round trip time, transmission cost. In [6], authors proposed a model for resource-task
mapping which could reduce execution cost and also designed a PSO based heuristic to allocate
tasks to resources. Both computation cost and data transmission cost are calculated by using the
workflow application. Authors compared results of heuristic against “Best Resource Selection”
(BRS) heuristic and found that PSO based task scheduling could result into three times cost
savings. In [7], authors compared three popular heuristic approaches namely PSO, GAs and
MPSO for efficient task scheduling in cloud environment. MPSO algorithm improved makespan
characteristics when compared with PSO and GA. In [8], authors proposed a hybrid of particle
swarm optimization and simulated annealing. CloudSim toolkit [add reference] was used to
implement and analyze the proposal. Population based metaheuristic (PBM) algorithm maximizes
resource utilization and minimizes makespan and demonstrating improvements upwards of 53%.
Hybrid meta-heuristic method proved to enhance the performance in CloudSim toolkit. In terms
of resources utilization and makespan Hybrid PSO implementation in [9] tried to balance the load

69
across the system and minimize the makespan. In [10], authors presented multi-objective PSO
based optimization algorithm for dynamic environment of clouds and optimize energy and
processing time. Proposed algorithm provides an optimal balance results for multiple objectives.
Experimental results illustrated that proposed methods (TSPSO) out-performed BRS and RSA.In
CA-PSO authors proposed Modified PSO by augmenting PSO with Cost-Aware Heuristic. CA-
PSO generates tasks schedules using a cost aware fitness function to quantify the cost of resource
usage along with fitness function for time cost to minimize processing time. Authors preferred to
optimize usage cost of resources if possible. In [11], a task-level scheduling algorithm Chaotic
Particle Swarm Optimization (CPSO) is based on chaotic sequence and inertia weight factor.
Chaotic sequence with high predictability improves the variety of solutions and its reliability
assures a good global convergence. It can optimize the cost of whole scheduling and overcome
the premature convergence of PSO algorithm to satisfy the market-oriented characteristic of cloud
workflow. The performance of CPSO is efficient and the cost is the lowest as compared with
other representative counterparts.
3.2 Ant Colony Optimization (ACO)
ACO is available as metaheuristic in Computer Science and Operation Search problems. This is a
probabilistic search method and can be applied to any problem which can be reduced to graphs
and searching optimal paths through graphs. Since its inception in 1995, ACO has gone through
several extensions and has seen wider acceptance in many of the engineering domains. ACO
simulates the searching behaviour of ants through ant-colony in search of foods. Some of the
popular extensions of ACO like, Elitist ant system, Max-min ant system (MMAS), Rank-based
ant system, Continuous Orthogonal Ant Colony (COAC) and Recursive ant colony optimization.
This section presents review of few proposals in ACO based scheduling solutions in cloud
environment.
In [12] authors established superiority of Ant Colony Optimization (ACO) algorithmsagainst
heuristic approaches like FCFS and RR. Performance of ACO in task scheduling is appreciable
when makespan of ACO based task scheduling is considered. In [13], Authors proposed ACO
based system that increases the profit of cloud service provider by outsourcing task to external
cloud when its resources are not enough. ACO outperforms the existing SLPSO (Self-adaptive
learning PSO) and provides optimized scheduling. The ACO based system decreases response
time (making it more responsive) and throughput as compared with existing SLPSO system. In
[14], authors proposed ACO-LB (ACO based Load Balancing)to target the imbalance of task
loads assigned to VMs during task scheduling. Heuristics based task scheduling schemes fails to
consider the load imbalance. Another issue with scheduling in cloud is dynamic arrival rate of
tasks. Solution proposed in this paper is able to adapt to dynamic environment of clouds. The
simulation based analysis established performance abilities of ACO-LB in Load Balancing. In
[15], MACOLB (definition) was introduced to balance system load and minimize the makespan
of given tasks so that performance of overall system can be improved. MACOLB achieved better
system load balance and reduced makespan as compared with ACO and MACO algorithms.
Simulation’s results shows that MACOLB algorithm outperforms MACO and ACO algorithms.
In [16] authors proposed a Load Shared Ant Colony Optimization (LSACO) to balance the load
across system. Proposal simulated across tasks and resources of varied types. Decision to share
load is based on storage requirements of tasks. The proposal successfully reduced response time,
improved system throughput and resource utilization. In [17], authors proposed a new constrained
based workflow scheduling algorithm to maximize reliability of workflow. In this paper authors

70
proposed three novel workflow scheduling heuristicswhich are based on Ant Colony System. The
aggregation of these heuristics minimized violations of reliability and constraints. Authors find
feasible schedules for time and cost considerations in resource selection by two heuristics and
other is used to enhance reliability. Simulation result on real world workflows shows that
proposed method increases feasibility ratio and reliability for various pairs of deadline and
budget. Authors also overcome instances of reduction in Grid profit by increasing feasibility
ratio.In[18], Authors proposed Energy Aware Ant Colony Optimization that reduces the energy
consumption for cloud datacenter and to find the hosts that consume least power based on their
CPU utilization and schedule tasks. Performance is compared with existing techniques. Proposed
method improves Make-span and conserves energy consumption in hosts up to 22%.
3.3 Artificial Bee Colony (ABC)
ABC was proposed by Karaboga in 2005.ABC was inspired by honey bees’ behaviour. ABC
simulates foraging behaviour of honey bees. Set of bees is called swarm and completes any tasks
by using social behaviour and cooperation. Bees are specified to be of three types namely,
employed bees, onlooker bees and scout bees. Employed bees search for food sources and
onlooker bees perform selection of good sources by using the information conveyed by employer
bees. Fitness function is specified to rank the quality of the food source. ABC uses fitness
function to rank any solution over the solution space. Higher fitness values ranks higher quality.
Scout bees are translated from a few employed bees, which abandon their food sources and search
new ones. Scout bees are promoted as employed bees, which search newer food sources. The
application of honey bees’ behaviour through ABC algorithms has been tested in scheduling in
many proposals and is reviewed in this section.
In [19], authors compared the performance of ABC optimization schemes with Evolutionary
Algorithms, PSO, and Differential Equations etc. for assessment of ABC to multi-dimensional
numeric problems. The behaviour of honey bees was imitated by using Artificial Intelligence.
Simulation results show that ABC is equally applicable to multidimensional engineering
problems. In [20] authors presented a new bee colony algorithm to schedule tasks on service
providers in Cloud environments. The proposed algorithm has better results of simulation in
terms of task execution time, waiting time and missed tasks.In [21] the Honey bee forage
mechanism was proposed for load balancing in cloud to utilize its resources in cloud and was
applied to optimize scheduling of Virtual Machine (VM) and reducing makespan, time a task
must be heldin a queue on VM. The proposed algorithm reduced makespan of processing time
and improves general turnout of process and priority based balancing.In [22] authors presented
anABC Algorithm that computes total finishing time, mean finishing time, load balancing time
after scheduling task. Authors considered time as main QoS factor. Algorithm produced best total
task finishing time, mean task finishing time and load balancing time as compared with ACO.In
[23], authors proposed Two-stage Artificial Bee Colony (TABC) algorithm for scheduling andre-
scheduling with new job(s) inserting. The algorithm is compared with six simple heuristics and
proposed hybrid heuristics. TABC adds crossover operator and critical path based local search for
solving flexible job-shop scheduling problem (FJSP) problem. The proposed algorithm obtains
better results in all 30 repeats than compared simple and combination eight heuristics in both
initializing scheduling and re-scheduling phase. Result shows TABC algorithm has very good
convergence and stability and to minimize makespan. In [24] proposed the hybrid algorithm
TABC a combination of artificial bee colony (ABC) and TabuSearch (TS) algorithm. TS based
strategy was embedded in TABC to solve the large-scale hybrid flow shop scheduling problem
with limited buffers and was tested on sets of instances with large scales that are generated based

71
on realistic production. The proposed algorithm balanced exploitation and exploration abilities of
the algorithms and TS-based local search is applied to the employed bees and onlookers with
different functions. In [25] authors compared the ABC algorithm with GA to block a solution
from allocating colossal tasks on a single machine and also to maintain balance of work allocation
among machines. Author found that the ABC performed better than GA to minimization of
maximal machine workload, represented by maximum runtime consumed by any machine and
symbolized by overall run time allocated to all the machines. It also reduced make span of tasks
as compared with Genetic.
3.4 BAT Optimization (BAT)
In [26] authors proposed a new metaheuristic approach and also included the good properties of
other approaches. The behavior of this optimization is based on echolocation behavior of BATs.
BAT uses a special type of SONOR called echo-locater to locate their hives, obstacles etc. BATs
generate a sound which echoes back and is used by BATs to locate the obstacles. Each sound
burst last 10-20 milliseconds but the waves have constant frequency between 25-150MHz. Micro
bats can generate 10-20 such sound burst and can be increased upto 100-200 burst per second,
when BATs chase any prey. Some bats have good eyesight and some have good smell senses.
The overall effort is to combine all these and obtain an effect which givesefficient prey detection
and a smooth navigation. Such echolocation function was associated with objective function to be
optimized and a new optimization algorithm called Bat Algorithm (BA) was proposed. BA was
composition of PSO and intensive local search controlled by loudness and pulse emission rate.
The proposal was compared with Genetic Algorithms (GAs) and PSO. PSO performed better in
terms of efficiency and accuracy than GAs, while BATalgorithm performed superior of all. BA
added the best feature of PSO and new feature of bats echolocation. In [27] authors introduced
chaos into BA to increase its global search mobility for robust global optimization. CBA selected
different chaotic systems to replace the parameters in BA and used these different methods for
potentially efficient alternatives to pseudo random sequences. The success rate of the CBAs
shown in Statistical results proved that the improved algorithms can improve the reliability of the
global optimality and quality of the results.In [28] presented a novel BCO based approach,
namely JDS-BC, to schedule jobs to CNs and replicate datafiles in SNs. JDS- BC focuses on two
independent matchmaking process and to minimized makespan and total transfer time
concurrently in system.
Table 1 Metaheuristic Scheduling Scheme Parametric Analysis
S.
No.
Constrained Goal Scheduling
Factor/Parameter
Task
Type
Tool Achieved Gaps
patched/
Contribution
[2] 1. Global
optimum.2.
Reasonable
number of
iterations
Based on varying
inertiaweight in range
[0.9,1.2]
I C
Language
1. ImprovedPSO
Performance
2. Inertia
Introduced
1. PSO without
inertia fails to
find the global
optimum within
4000 iterations
[3] 1. Minimize
processing time
2. Maximize
resource usage
1. Cost-Aware Based
PSO
2. Scheduling
Algorithm
I 1.MatLab2.
CloudSim
1. Task
scheduling 2.
Global optimal
solution
1. Earlier PSO
failed to
consider the
resource usage
cost
[4] 1. Deadline - 1. PSO_HOM
(PSO based
heuristic)
W 1.
CloudSim
1. Minimize
Execution Cost
2. Meeting
1.Performanc
e Comparison
with other

72
Defined
Deadline
Optimization
Strategies and
PSO
2. Find the
Data Transfer
Cost
between Data
Centers sothat
VMs can be
deployed in
different
regions.
[5] 1.
Reliability2.
Availability
- 1. Load Balancing
2. Mutation
3. PSO
I 1.
CloudSim
1.Minimize cost
2. Minimize RTT
3. Minimize
Execution Time
4. Minimize
Transmission
Time
5. Load
Balancing
6. Improved
Reliability
7. Resource
Availability
1. Only
Uniform VM
Characteristic
have been
considered.
[6] - 1.
Computation
Cost
2. Data
Transmission
Cost
1. PSO based
Heuristic
W 1. Java
based
JSwarm
package
1. Three times
Cost Savings
as compared to
BRS
1. Integrate
PSO based
heuristic into
Brain
Imaging
Analysis
[7] - 1. Optimal
schedule
1. Modified PSO I 1.
Eucalyptu
s Cloud
1. Reduced
Makespan
1. Evaluated
multiple
System
Performance
Parameters
[8] - 1. Minimize
makespan
2. Increase
Resource
Utilization
1. Population-
Based
Metaheuristic
(PBM) algorithm
2. Simulated
Annealing
I 1.
CloudSIm
1.
Improvement
upwards of
53% in
makespan
2. Resource
Utilization
1.
Implementati
on of a multi-
objective
solution
[9] - 1. Load
Balancing
1. Hybrid PSO I 1.
Cloudsim
1.Minimize
Makespan
2.Improved
Resource
Utilization
1. Integrate
HPSO inmap-
reduce
framework
for Task
Scheduling
[10] - 1. Energy
Efficiency
1. Nested Particle
Swarm
Optimization(TSP
SO)
W 1.
CloudSim
1. Optimization
Energy
Consumption
2. Overall
Computation
Time
1. Bandwidth
usage
Analysis2.
Load
Balancing
3. Cost
Etc.
[11] - 1. To reduce
the cost of
Workflow
system
1.Chaotic Particle
Swarm Optimization
(CPSO) algorithm
W 1.
Workflow
Sim
1. Improves
the diversity of
solutionsand
1. Experim
ent with
more

73
good global
convergence
than 4
VMs not
be
conducte
d
2. QoS
constrain
ts are not
consider
ed
[12] - 1. Minimize
Makespan
1. Ant Colony
Optimization
(ACO)
I 1.
CloudSim
1. Minimize
Makespan
1.Task
Precedence
of Tasks
2. Load
Balancing
3.Compared
with other
Metaheuristic
s
[13] 1. Maximize
Profit
2. QoS
1. Ant Colony
Algorithm (ACO)
I 1.
CloudSim
1.High
Throughput
2.Improved
Economy by
0.26-11.58%
1. Swarm
Intelligence
and ACO
Hybridization
[14] - 1. Load
Balance of
VMs
1. Load Balancing
based
Ant Colony
Algorithm (ACO-
LB)
W CloudSim 1.Reduced
Makespan
2.Increse
Utilization
3.Balance Load
1.Different
types of
workflows
were used to
verify the
reliability of
the ACO-LB
algorithm
2.Cost
Problem
[15] - 1. Load
Balancing
1. Modified Ant
Colony
Optimization for
Load Balancing
(MACOLB)
I 1.
CloudSim
1.Minimize
Makespan
2.Optimal
Resource
Allocation
3.Balance Load
1.Cost
2.Task Types
[16] 1. Memory
Constrained
- 1. Load Shared
Ant Colony
Optimization
(LSACO)
I Java 1.Reduces
Response Time
2.Increases
Resource
Utilization
3.Incresed the
no. of tasks
scheduled
1. Considered
ETC matrix
as Parameter
to get the
Perfect Wait
Time
And
Response
Time
[17] 1.
ReliabilityDr
iven
Scheduling
- 1. Constrained
Workflow
Scheduling
(CWS) algorithm
W GridSim 1.Maximize the
reliability of
workflow
Execution
2. Meet user
defined Budget
and Deadline
3. Reduces the
Grid Profit
1. Variance in
Load of
Resources can
change the
Execution
Time and
Cost of
activities

74
Loss by
increasing the
Feasibility
Ratio
[18] - 1. Energy
Efficiency
1. Energy-
AwareAnt Colony
Algorithm (ACO)
I CloudSim 1.Improves
Energy
Savings up to
22%
2.Improves
Makespan
1. Implement
IO Bound and
Dynamic
Workload
2. Energy
Management
in the Cloud
[20] 1. Tasks
Scheduling
1. New Bee
Colony Algorithm
W Cloudsim 1.Reduced
Task
Execution
Time
2.Reduced
Waiting Time
1. Task
Scheduling
[21] - 1. Load
Balancing
1. Honey Bee
Behavior
Galvanized Load
Balancing
Algorithm
I CloudSim 1.Reduce
Makespan
2.
ReducedRespo
nse Time
1. Considered
alternative
QoS factors
of Tasks
[22] - 1. Task
Scheduling
1. ABC I CloudSim 1. Minimize
Makespan
2. Mean
Makespan
3. Load
Balancing
Time
1. Improved
Datacenter
Efficiency by
adding more
Parameters
[23] - 1. Flexible
Job-Shop
scheduling
Problem
(FJSP)
1. Two-stage
Artificial Bee
Colony (TABC)
I C++ 1. Minimize
Makespan
1. Additional
local
Search
Heuristics
[24] - 1. Large-scale
Hybrid Flow
Shop
scheduling
problem
with Limited
Buffers (HFS-
LB)
1. Hybrid Artificial
Bee Colony
(HABC)
I C++ 1. High
Exploitation
and
Exploration
capabilities in
solving Large-
scale HFS-LB
1.
EnhancedBala
nce between
Exploitation
and
Exploration
2. Reliability
[25]
[26]
[27]
-
-
-
1. Reduce
Make-span
1. Improved
Algorithmic
Convergence
1. Increase
1. Artificial Bee
Colony
(ABC)
1. BAT
Algorithm
(BA)
1. Chaos based
I
I
I
MATLAB
MATLAB
-
1.Minimize
Makespan
2.Minimize
Workload on
any VM
3. Minimize
Workload on
all machine
1. Accuracy
2. Efficiency
1. Improves
1. This
process can
be analyzed
with Parallel
Algorithms

75
Global
Search
Mobility
BAT
Algorithm
(CBA)
Reliability
2. Enhance
the quality
of results
[28] - 1. Job Data
Scheduling
1. Job Data
Scheduling using
Bee Colony (JDS-
BC)
I GridSim 1. Makespan
2.Total Data
file Transfer
Time
1. Energy
Efficiency of
Grid
Resources 2.
Cost of
System
Maintenance
*I= Independent, W = Workflows and B = Batch Jobs
3. GENETIC ALGORITHM (GA) BASED SCHEDULING
GA is proved to be an effective method for searching over complex and difficult search space.
GAs is collection algorithms which uses genetic operators like crossover, mutation and selection
etc. GAs integrates other problem solving techniques with genetic operators. Combing the other
approach with genetic operators generates Genetic Algorithms. GAs suffers from premature
convergence of and may produce only local optimum values. This limitation of GA guides for
hybridization of GAs with other techniques like metaheuristics and propose context specific
hybrid GAs.
4.1 Current and Historic GAs Proposals:
In [29] authors proposed a genetic algorithms- (GAs-) based approach that effectively solved the
problems of multi objective optimization. A genetic algorithms-based approach can be used to
make decisions on how to assign tasks to cloud nodes and how to migrate task in pervasive
clouds. Allocation strategies are used for optimized solutions and system can make use of the
chosen allocation scheme to execute user tasks. The GA based approach provides the best
solution for task migration in cloud environment.In [30] authors have proposed a GAbased
scheduling heuristic for solving budget and deadline constrained optimization problems and
scheduling workflow applications by either minimizing the monetary cost, execution time while
meeting users’ budget constraint and deadline constraints. Approach was compared with other
heuristics on two workflows; balanced and unbalanced structures and results provedthat GA
performed better in handling a complex workflow structure. A Hierarchic Genetic Scheduler
(HGS-Sched), a 2-level hybrid of Genetic Algorithm and Tabu Search (GA + TS) strategy for
improving the efficiency of the single-population genetic-based schedulers in the dynamic grid
environment was proposed in [31]. The HGS-Sched enabled a simultaneous search of solution
space by many small dependent populations. In comparison with mono-population and hybrid
genetic-based schedulers the proposed algorithm proved highly effective and resulted in fast

76
reductions in makespan and flowtime.In [32] authors proposed Multiple Priority Queues Genetic
Algorithm (MPQGA) which incorporated a genetic algorithm (GA) approach which is suitable
for scenario of Directed Acyclic Graph (DAG) scheduling designed with Crossover, Mutation and
Fitness function. Proposal exploited the improvement of both Evolutionary-based and Heuristic-
based algorithms while avoiding their drawbacks. Proposed MPQGA algorithm outperformed two
non-evolutionary heuristics and a random search method in terms of schedule quality and covered
a larger search space without incurring high computational cost than deterministic scheduling
approaches.In [33] authors proposed a GA based load balancing strategy to balance the load and
minimizing the makespan of given tasks provide an efficient utilization of resource in Cloud
Computing infrastructure. In comparison with FCFS and RR and a local search algorithm
Stochastic hill climbing, proposed GA algorithm outperform in terms of load balancing but also
guarantees the QoS requirement of customer job. In [34] authors proposed Task scheduling
algorithm using Improved GA. Authors compared output of improved GA with existing GA and
found that improved GA reduced overall execution time of the tasks and improves utilization of
resources.
Table 2 Genetic Algorithms Scheduling Schemes Parametric Analysis
S.
No.
Constraint Goal Scheduling
Factor/Param
eter
Task
Type
Tool Achieved Gaps Patched/
Contribution
[29] - 1.
Optimized
Tasks
Allocation
Scheme
1. GA based
Approach
I CloudSim 1. Fastest
Execution
Efficiency
2. Improve
Task allocation
and task
migration
efficiency
1. Dynamic
Decision according
to Cloud
Environment
2. User’s tasks
Status
3. Context-
Awareness
Framework
[30] 1. Deadline
2. Budget
- 1. GA
basedApproach
W GridSim 1. Reduced
execution Cost
2. Reduced
Time
1. Enhanced
scheduling
algorithm by
using different
service models
2. Dynamic Data-
driven workflow
models
[31] - 1.
Improving
Effectiven
ess of
Single-
Populatio
n GA
based
schedulers
1. Hierarchic
Genetic
Scheduler
(HGS-Sched)
Iand
B
1. Sim-G-
Batch grid
1. Minimized
makespan
2. Flowtime
1. Improvement of
effectiveness of
Pareto front
underMulti
Objective
Grid Scheduling
[32] - 1. Task
Schedulin
g on
Heterogen
eous
Computin
g Systems
using
1. Multiple
Priority
Queues GA
(MPQGA)
W C# 1. Reduced
Makespan
2. Improve
Speedup
3. Efficiency
1. Considered
More Task graphs
2. MoreProcessors
3. Variable
Degrees of
Heterogeneity
among Processors
and Subtasks

77
Multiple
Priority
Queue
[33] - 1. Load
Balancing
1. GA I Cloud
Analyst
1.Load
Balancing
2.Minimizing
Make-span
1. Variation of the
Crossover and
Selection
[34] - 1. Task
Schedulin
g
1. Improved
GA
I CloudSim 1. Reduced
Makespan
1. Reduced
Makespan
5. GREEDY APPROACH BASED SCHEDULING SCHEMES
Greedy approach is powerful and time tested approach. Several problem domains like shortest
paths, spanning trees, activity selection etc. are popularly solved by using greedy based
algorithms. Greedy algorithms are applied to problems with certain features. First requirements
problem substructure, i.e., problem can be divided into sub-problems. Another requirement is that
there exists a greedy choice property. Greedy choice property lets next local best step. In general,
one of the sub-problems turns empty and other one non-empty. All in all there is only one sub-
problem plus last local best step. Several proposals used greedy approach to solve scheduling
problems in cloud environment.
In [35] authors proposed Greedy Method based Algorithm for cloud environment for improving
makespan performance and user satisfaction. In [36] authors presented a greedy based approach
to optimize scheduling which addressed the major challenges in cloud scheduling. Two level of
job scheduling and rescheduling was performed by considering parameters like deadline, cost,
profit, response time, waiting time and turnaround time etc. Greedy technique was used to select
resources on the basis of task constraints. Re-Scheduling is based on cost/profit parameters. In
comparison with existing algorithms like FCFS, Round Robin (RR) and Priority algorithms, the
proposed algorithm improved the efficiency. Greedy technique is used to select resources on the
basis of task constraints and improved the efficiency and resource utilization. In [37] authors used
greedy approach to select the appropriate resource and minimize their execution cost in
comparison of sequential approach and also reduced execution time of tasks. In [38] authors
presented greedy based task scheduling to improve turnaround time performance and reduce cost
incurred for user submitted tasks in particular time slot. Proposed algorithm selected the resource
on the basis of its cost and turnaround time. Priority formula is used for task selection. Proposed
algorithm gives better results over sequential scheduling. In [39] proposed Min-Max algorithm to
verify its validity with Min-Min and Max-Min. It puts together small, big tasks and makes good
use of time of greedy strategy in order to solve problem of load imbalance. The Min-Max results
proved that saved up to 9% overall execution time compared with Min-Min and improved the
utilization rate of the entire system compared with Max-Min and the total completion time and
average response time reduced by 7% and 9% respectively. In [40] proposed task scheduling
algorithm which respects the constraints specified for the said task. Authors adopted greedy
method like activity selection using greedy approach. The proposed algorithm reduced turnaround
time and cost characteristics of each task submitted.

78
Table 3 Greedy Algorithms based Scheduling Schemes Parametric Analysis
Sr.
No.
Constraine
d
Goal Scheduling
Factor/Param
eter
Task
Type
Tool Achieved Gaps
/Contribution
[35] - 1. Job
Scheduling
1. Greedy-Based
job scheduling
I CloudSim 1.Reduced
Execution
Time
2.Improved
QoS
1. Improvement of
the proposed algorithm
in order to gain more
fairness
[36] 1. Deadline
2. Cost
- 1. Greedy
Algorithm
I CloudSim 1. Improved
Efficiency
2. Resource
Utilization
1. More parameters for
evaluation
[37] - 1. Task
Scheduling
2.
Resource
Allocation
1. Greedy
algorithm
W Eucalyptus 1. Reduced
Execution
Time
2. Cost
1. Cost based
scheduling policy can
be implemented
[38] - 1. Task
Scheduling
1. Greedy
Algorithm
W CloudSim 1. Reduced
Task
Completion
Time
2. Execution
Cost
1. Add location
parameter with the
existing Deadline and
Cost
[39] - 1. Task
Scheduling
1. Min-Max I CloudSim 1. Improves
Utilization
2. Task
Completion
Time
3. Average
Response
Time
[40] - 1. Task
Scheduling
1. Dynamic
Optimization
Algorithm based
on Greedy
Approach
I CloudSim 1. Reduced
Turnaround
Time
2. Cost
1. Type of task
2. Task Length could
be taken into account
6. HEURISTIC BASED SCHEDULING
Heuristic Techniques is some practical and non-guaranteed method of problem solving in the
field of science, technology etc. Heuristic technique may not guarantee the optimal solution but
may prove helpful for immediate solution of the problem at hand. Heuristics are generally derived
from rule-of-thumb, common sense, intuitive guess or educated judgements. Several proposals
have used Heuristic based Approach for scheduling problems in cloud environments.
Authors in [41] proposed three different heuristics to assess importance of task scheduling in
MapReduce. The parameter of importance is makespan. The scheduling problem was modeled as
two-stage hybrid flowshop scheduling problem with schedule-based setup times. Jobs and tasks
are considered two distinct entities. Two of the heuristics proposed in the paper are classified as
job sequencing heuristics and one as task sequencing heuristic. Authors incorporated data locality
of tasks in scheduling of tasks. Authors also proposed a lower bound of the makespan. The
performance of heuristics depends upon number of jobs and cluster setups. Simulation based
study compared performance of three heuristics and data dependent modeling is proved effective.
In [42] authors proposed a heuristic based on cost as QoS parameter for optimization. The

79
experiment was simulated using cloudsim3.0 toolkit and shown improved cost effective
performance. QoS driven algorithm considered execution time, load and latency as important
parameters. The performance was compared against FCFS. In [43] authors proposed a scheduling
heuristic to accommodate task priority, task length and user satisfaction. The historic solutions
like min-min don’t consider the task priority and thus ignored the satisfaction of user submitting
the tasks to cloud environment. Authors proposed an approach to assign credits to tasks on the
basis of task lengths and task priority. This has helped the cloud service providers to deal with
tasks with same priority. Task credits obtained in the proposal are unique in the set of the task
submitted for perusal. The performance of the credit based task scheduling was compared with
that of scheduling algorithm with priority only and length only credits. In [44] authors presented a
Maximum Effective Reduction (MER) algorithm which trades minimal increase in makespan
against maximal resource usage reduction on the schedules generated by any given scheduling
algorithm. Novelty of the approach is near-optimal trade off point determination. The tradeoff
point consideration leads to reduced resource provisioning, improving resource utilization and
reduced energy expenses. Another broad applicability of MER lies in the fact that it can be
applied to any scheduling environment dealing with scientific workflows. MER works by
consolidating tasks from over provisioned under- utilized resource to other resources and identify
to reduce resource provisioning. The performance of MER was compared with Critical Path First
(CPF), Critical-Path-On-Processor (CPOP) and Dynamic Critical Path (DCP). The exhaustive
simulation reports about 52% of resource usage reduction against 10% increment in makespan. In
[45] the authors reviewed several taxonomies for classification of challenges in Work Flow
Scheduling schemes. The taxonomies are specified on the basis of QoS system functionality and
system architecture. Authors also presented schematic on selecting the proposals suitable for
inclusion in the review. Authors also identified future directions on the basis of schematic
pursued. In [46] authors proposed a heuristic called Cost-Greedy Price-Adjusting (CGPA) for
achievement of multiple objectives namely; System throughput, Total Cost and Profit Fairness.
The heuristic benefits both grid users and providers. CGPA lowers the Fairness Deviation of
Profits which is not generally taken care of by other popular algorithms. The algorithms also
specified by the authors and simulation based performance of CGPA has been compared with
four other popular heuristics. A fair amount of experimentation was carried out and leads to
achievement of all the objectives. In [47] proposed a Constrained Workflow Scheduling (CWS)
algorithm that is based on Ant Colony Optimization (ACO) to maximize the workflow execution
reliability while respecting the user-defined deadline and budget. The author proposed three novel
heuristics which are adaptively selected by ants. Two of them are employed to find feasible
schedules and the other is used to enhance the reliability. Thecapability of CWS to find feasible
schedules and compared with greedy approach toward optimization of time, reliability and cost of
activities. The proposed algorithm has improved the feasibility ratio for various constraints,
reliability and reduced the Grid profit loss.In [48] authors proposed community aware scheduling
algorithm (CASA) that is one of the best dynamic scheduling approach are designed to efficiently
distribute jobs amongst participating nodes. In cloud environment the CASA is comprised of two
phase scheduling that is set of heuristic sub-algorithms to achieve optimized scheduling
performance The development made by an algorithm in this field of cloud has made a great
contribution of task scheduling. In a decentralized scheduling manner without requiring detailed
real-time processing information the use of CASA can lead to a 30%–61% better average job
slowdown and a 68%–86% shorter average job waiting time compared with two reference
conventional scheduling scenarios. The proposed community-aware scheduling algorithm
(CASA) has improved performance when it works with faster response time information system,
as compared with a slow reaction speed information system with integrated global
knowledge.In[49] authors proposed a new Pareto-based multi-objective workflow scheduling

80
algorithm as an extension to an existing state-of-the-art heuristic capable of computing workflow
schedules representing tradeoffs between energy consumption and makespan. This approach
captures the real behavior of energy consumption that is based on empirical models in
heterogeneous parallel systems. The proposed algorithm compared with a classical mono-
objective scheduling heuristic and state-of-the-art multi-objective optimization algorithm and
finding it computes better or similar results in different scenarios. The MOHEFT compared with
HEFT and green HEFT, an ad-hoc greedy algorithm under different experimental configurations
and observes it reduced the energy consumption by up to 34.5% with a slight increase of 2% in
the makespan. In[50] proposed a Cloud-aware data intensive workflow scheduling system
considered the QoS constraints of each workflow and improves the system performance by using
Cloud resources by realistic scientific workflows. The proposed algorithm partitioned a workflow
into sub-workflows to minimize data dependencies among the sub-workflows to distribute on
volunteer resources. The proposed provisioning algorithm (VOLNT-SPOT) compared the
performance with the system that contains just volunteer resources (VOLNT) as it improves the
system performance by increasing the percentage of workflows that meet the deadline with a
factor of 75% in average with respect to the execution of workflows on the volunteer resources.
Two Policies are considered in deploying public Cloud resources significantly to increase the
percentage of workflow in which first policy considers 23% saving in cost average than the
second policy while the second strategy just considers the relaxed deadline by a larger margin.In
[51] proposed a hybrid approach for tasks scheduling in Heterogeneous Distributed Computing
(HySARC2) and used hierarchical clustering of the available resources into groups. Authors
considered different DAG scheduling strategies for independent task. Proposed algorithm is
suitable for Heterogeneous Distributed Computing, especially for modern High-Performance
Computing (HPC) systems in which IO and computational intensive requirements are modeled.
The HySARC2 algorithm more efficient and exact structure of resources because it analyzed the
available resources and groups them into clusters (resource aware algorithm) and prerequisite
different groups of similar tasks to different clusters of resources and then scheduled the tasks in
each cluster of resources. It improves workload on the resources in cloud environment with
respect to load-balancing, cost savings, dependency assurance for workflows.
Table 4 Heuristic based Scheduling Schemes Parametric Analysis
Sr.
No.
Constrai
ned
Goal Scheduling
Factor/Parameter
Task
Type
Tool Achieved Gaps/Cont
ribution
[41] - 1. Job
Scheduling
2. Task
Scheduling
1. Job Sequencing
Heuristic
2. Task
Sequencing
Heuristic
I MapReduce 1.Lower bound
on Makespan
2. Data
Dependency
Established
3. Cluster Set
updependency
Established
1. Data
dependency
incorporate
d
2. Lower
bound on
Makespan
[42] - 1. Cost
2.
Execution
Time
1. Task Scheduling I CloudSim 1. Cost based
Analysis
1. More
parameters
for
evaluation
can be
added
[43] - 1. Task
Scheduling
2.
Makespan
1. Credits based
Scheduling
I CloudSim 1. Improved
Makespan
Characteristics
2. Credit based on
1. Varying
the VM
characterist
ic could

81
Length and
priority of tasks
have been
considered
[44] - 1. Schedule
Optimizatio
n
2. Reduced
resource
Provosioni
ng
3.
Improved
Resource
Utilization
1. Maximum
Effective
Reduction
Heuristics
W CloudSim 1. Reduced
Resource
Provisioning
2. Improved
Utilization
3. Task
Consolidation
1.
Optimized
schedules
generated
by other
heuristics
[45] - 1. Cost
Aware
heuristic
taxonomies
2. Future
Directions
1. Taxonomy for
cost aware Work
flow Scheduling
W - 1. Taxonomy
considering QoS
2. System
Functionality 3.
System
Architecture
-
[46] - 1. Profit
Fairness
2. Total
Cost
3. System
Throughput
1. Cost Greedy
Price Adjusting
(CGPA) Heuristic
W - 1. Lower Fair
Profit Deviation
2. Improved
Throughput
1. Fair
Allocation
in terms of
resource
and cost
[47] 1.
Deadline
2.
Budget
- 1. Constrained
Workflow
Scheduling
(CWS) algorithm
W Grid
Sim
1. Increases
feasibility ratio
2. Reliability for
various pairs of
deadline and
budget
1.Other
kinds
of
constrained
workflow
scheduling
can be
investigate
d.
2. The
variation in
load of
resources
can change
the
executionti
me and cost
ofactivities.
[48] 1.
Decentraliz
ed
Dynamic
Scheduling
1. Community-
Aware Scheduling
Algorithm (CASA)
I MaGate
Sim
1. Faster
Response Time
2. Improvement
in both the
average job
slowdown and
average job
waiting time
3. Improves
Resource
Utilization
1.
Implementa
tion
of CASA
within the
context of
current
developme
nts in
cloud
computing
[49] - 1.
Makespan
2. Energy
Efficiency
1. Multi-Objective
Heterogeneous
Earliest Finish
timealgorithm
(MOHEFT)
W POV-Ray 1. Reduced
Energy
2. Makespan
1. Optimize
Economic
Cost
2. Extend
MOHEFT

82
to consider
multi-
tenantcom
mercial
cloud
systems.
[50] 1. QoS2.
Deadline
- 1. Cloud-Aware
Data
Intensive
Workflow
Scheduling
W Cycloid
Grid
1. Deadline
2. Cost
1.
Volunteer
resources
2.
Decreases
the number
of Amazon
spot
instances
[51] - - 1. Hybrid approach
for tasks
Scheduling in
Heterogeneous
Distributed
Computing
(HySARC2)
W Cloud
Sim
- 1. Dynamic
behavior of
the
resources
and
algorithms
2. Allow
Tasks
Preemption
7. CONCLUSION
This paper presents a review of Task and Job Scheduling schemes. A novel taxonomy is proposed
in the paper. Schemes falling under Goal Oriented Task Scheduling (GOTS) schemes give service
providers a fair chance to apply approach and schedule the tasks and resources that can generate
maximum possible economic gains, while using least resource provisioning. Using low resource
provision allows providers to use their resources at possible fullest and trading makespan with
marginal increase only. Certainly GOTS is provider friendly policy. At the same time Constraint
Oriented Task Scheduling (COTS) allows the user to specify their priority and importance. To
ensure that desired level of service requests are met by service providers, SLAs are used. Users
can specify their QoS parameters and other requirements. This paradigm gives providers an
opportunity to negotiate the agreements at their own costs. COTS allow users to enjoy the
flexibility and providers to earn higher. Table 1, 2, 3 and 4 gives a tabular glimpse of recent
trends in scheduling schemes for cloud environment. With extensive review of recent proposals in
scheduling schemes, we concludes that a lot more dimensions are yet to be explored in
scheduling, like data dependency, two stage optimization of already generated schedules, batch
jobs, Workflow technologies etc.
REFERENCES
[1] http://www.swarmintelligence.org/tutorials.php, accessed on May 14, 2015.
[2] Y. Shi and R. Eberhart, “A modified particle swarm optimizer”, Evolutionary Computation
Proceedings, 1998. IEEE World Congress on Computational Intelligence, The 1998 IEEE
International Conference on. IEEE, (1998).
[3] Zhao, G. (2014). ‘Cost-Aware Scheduling Algorithm Based on PSO in Cloud Computing
Environment’, International Journal of Grid and Distributed Computing. 7, 1, .33-42

83
[4] Rodriguez, M. A., &Buyya, R. (2014). Deadline based resource provisioning and scheduling
algorithm for scientific workflows on clouds. Cloud Computing, IEEE Transactions on, 2(2), 222-
235.
[5] Awad, A. I., El-Hefnawy, N. A., &Abdel_kader, H. M. (2015). Enhanced Particle Swarm
Optimization for Task Scheduling in Cloud Computing Environments. Procedia Computer Science,
65, 920-929.
[6] Pandey, S., Wu, L., Guru, S. M., &Buyya, R. (2010, April). A particle swarm optimization-based
heuristic for scheduling workflow applications in cloud computing environments. In Advanced
Information Networking and Applications (AINA), 2010 24th IEEE International Conference on (pp.
400-407). IEEE.
[7] Abdi, S., Motamedi, S. A., &Sharifian, S. (2014). Task scheduling using Modified PSO Algorithm in
cloud computing environment. In International Conference on Machine Learning, Electrical and
Mechanical Engineering (pp. 8-9).
[8] Hussein, S., Robert, C., Alam, M. (2014). ‘Cloudlet Scheduling with Population Based Meta
Heuristics’. Journal of Computer Science and Management Research. 1, 3, 348-351
[9] 1gomathi, B., Krishnasamy, K. (2013). ‘Task scheduling algorithm based on hybrid particle swarm
optimization in cloud computing environment’. Journal of Theoretical and Applied Information
Technology. 55, 1, 33-38.
[10] Jena R. K. (2015). Multi objective Task Scheduling in Cloud Environment Using Nested PSO
Framework, 21 Aug, 2015.
[11] Li, X., Xu, J., & Yang, Y. (2015).‘A chaotic particle swarm optimization-based heuristic for market-
oriented task-level scheduling in cloud workflow systems’. Computational Intelligence and
Neuroscience, 2015, 81-92.
[12] Tawfeek, M., El-Sisi, A., Keshk, A. E., &Torkey, F. (2013, November). Cloud task scheduling based
on ant colony optimization “, The International Arab Journal of Information Technology, Vol. 12, No.
2, March 2015 pp.129-137.
[13] Anuradha, M., &Selvakumar, S. ACO Based Task Scheduling Algorithm for Hybrid Cloud.
International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE) ISSN:
0976-1353 Volume 13 Issue 1 –MARCH 2015
[14] Shengjun, X., Mengying, L., Xiaolong, X., Jingyi, C. (2014). ‘An ACO-LB Algorithm for Task
Scheduling in the Cloud Environment’. Journal of Software. 9, 2, 466-473
[15] Keshk, A. E., El-Sisi, A., Tawfeek, M. A., &Torkey, F. A. (2013). Intelligent strategy of task
scheduling in cloud computing for load balancing. International Journal of Emerging Trends and
Technology in Computer Science, 2(6), 12-22.
[16] Kokilavani, T., &Amalarethinam, D. D. G. (2012). Memory Constrained Ant Colony System For
Task Scheduling In Grid Computing. International Journal of Grid Computing & Applications
(IJGCA) Vol, 3.
[17] Kianpisheh, S., Charkari, N. M., &Kargahi, M. (2016). Reliability-driven scheduling of time/cost-
constrained grid workflows. Future Generation Computer Systems, 55, 1-16.
[18] Linda J and Ananthanarayana V.S., “Energy-aware Task scheduling in cloud using Ant-colony
Optimization”, Proceedings of International Conference on Emerging Research in Computing,
Information, Communication and Applications (ERCICA-14)
[19] Karaboga, D., &Basturk, B. (2008). On the performance of artificial bee colony (ABC) algorithm.
Applied soft computing, 8(1), 687-697.
[20] Navimipour, N. J. Task scheduling in the Cloud Environments based on an Artificial Bee Colony
Algorithm. Proceedings of 2015 International Conference on Image Processing, Production and
Computer Science (ICIPCS'2015) Istanbul (Turkey), June 3-4, 2015 pp. 38-44
[21] Monika Rathore, Sarvesh Rai, NavdeepSaluja, “Load Balancing of Virtual Machine Using Honey Bee
Galvanizing Algorithm in Cloud”, International Journal of Computer Science and Information
Technologies, Vol. 6 (4) , 2015, 4128-4132
[22] Kumar, R. S., &Gunasekaran, S. (2014). Improving Task Scheduling in Large Scale Cloud
Computing Environment using Artificial Bee Colony Algorithm. International Journal of Computer
Applications, 103(5).

84
[23] Gao, K. Z., Suganthan, P. N., Chua, T. J., Chong, C. S., Cai, T. X., & Pan, Q. K. (2015). A two-stage
artificial bee colony algorithm scheduling flexible job-shop scheduling problem with new job
insertion. Expert Systems with Applications, 42(21), 7652-7663.
[24] Li, J. Q., & Pan, Q. K. (2015). Solving the large-scale hybrid flow shop scheduling problem with
limited buffers by a hybrid artificial bee colony algorithm. Information Sciences, 316, 487-502.
[25] Muthiah, A., &Rajkumar, R. (2014). A Comparison of Artificial Bee Colony algorithm and Genetic
Algorithm to Minimize the Makespan for Job Shop Scheduling. Procedia Engineering, 97, 1745-1754.
[26] X. S. Yang, A New Metaheuristic Bat-Inspired Algorithm, in: Nature Inspired Cooperative Strategies
for Optimization (NISCO 2010) (Eds. J. R. Gonzalez et al.), Studies in Computational Intelligence,
Springer Berlin, 284, Springer, 65-74 (2010).
[27] Gandomi, A. H., & Yang, X. S. (2014). Chaotic bat algorithm. Journal of Computational Science,
5(2), 224-232.
[28] Taheri, J., Lee, Y. C., Zomaya, A. Y., & Siegel, H. J. (2013). A Bee Colony based optimization
approach for simultaneous job scheduling and data replication in grid environments. Computers &
Operations Research, 40(6), 1564-1578.
[29] Zhang, W., Tan, S., Lu, Q., Liu, X., & Gong, W. (2015). A Genetic-Algorithm-Based Approach for
Task Migration in Pervasive Clouds. International Journal of Distributed Sensor Networks.
[30] Yu, J., &Buyya, R. (2006). Scheduling scientific workflow applications with deadline and budget
constraints using genetic algorithms. Scientific Programming, 14(3-4), 217-230.
[31] Joanna Kołodziej a, Samee Ullah Khan, “Multi-level hierarchic genetic-based scheduling of
independent jobs in dynamic heterogeneous grid environment”
[32] Xu, Y., Li, K., Hu, J., & Li, K. (2014). A genetic algorithm for task scheduling on heterogeneous
computing systems using multiple priority queues. Information Sciences, 270, 255-287.
[33] Dasgupta, K., Mandal, B., Dutta, P., Mandal, J. K., & Dam, S. (2013). A genetic algorithm (ga) based
load balancing strategy for cloud computing. Procedia Technology, 10, 340-347.
[34] Kumar, P., Verma, A. (2012). ‘Independent Task Scheduling in Cloud Computing by Improved
Genetic Algorithm’. International Journal of Advanced Research in Computer Science and Software
Engineering. 2, 5, 111-114
[35] Li, J., Feng, L., & Fang, S. (2014). An greedy-based job scheduling algorithm in cloud computing.
Journal of Software, 9(4), 921-925.
[36] Selvi, S., Maheswari, R., &Kalaavathi, B. Deadline–Cost Based Job Scheduling Using Greedy
Approach in a Multi-Layer Environment.
[37] Singh, L. S. V., & Ahmed, J. A GREEDY ALGORITHM FOR TASK SCHEDULING &
RESOURCE ALLOCATION PROBLEMS IN CLOUD COMPUTING. International Journal of
Research & Development in Technology and Management Science–Kailash, 21(1).
[38] Parikh, S., & Sinha, R. (2013). Double Level Priority based Optimization Algorithm for Task
Scheduling in Cloud Computing. International Journal of Computer Applications, 62(20).
[39] Zhou, Z., & HU, Z. (2014). Task scheduling algorithm based on greedy strategy in cloud computing.
Open Cybernetics &Systemics Journal, 8, 111-114.
[40] Gu, J., Hu, J., Zhao, T., & Sun, G. (2012). A new resource scheduling strategy based on genetic
algorithm in cloud computing environment. Journal of Computers, 7(1), 42-52.
[41] Li, X., Jiang, T., & Ruiz, R. (2016). Heuristics for periodical batch job scheduling in a MapReduce
computing framework. Information Sciences, 326, 119-133.
[42] Bansal, N., Maurya, A., Kumar, T., Singh, M., & Bansal, S. (2015). Cost performance of QoS Driven
task scheduling in cloud computing. Procedia Computer Science, 57, 126-130.
[43] Thomas, A., Krishnalal, G., & Raj, V. J. (2015). Credit Based Scheduling Algorithm in Cloud
Computing Environment. Procedia Computer Science, 46, 913-920.
[44] Lee, Y. C., Han, H., Zomaya, A. Y., &Yousif, M. (2015). Resource-efficient workflow scheduling in
clouds. Knowledge-Based Systems, 80, 153-162.
[45] Alkhanak, E. N., Lee, S. P., & Khan, S. U. R. (2015). Cost-aware challenges for workflow scheduling
approaches in cloud computing environments: Taxonomy and opportunities. Future Generation
Computer Systems.

85
[46] Xu, H., & Yang, B. (2015). An incentive-based heuristic job scheduling algorithm for utility
grids.Future Generation Computer Systems, 49, 1-7.
[47] Kianpisheh, S., Charkari, N. M., &Kargahi, M. (2016). Reliability-driven scheduling of time/cost-
constrained grid workflows. Future Generation Computer Systems, 55, 1-16.
[48] Huang, Y., Bessis, N., Norrington, P., Kuonen, P., &Hirsbrunner, B. (2013). Exploring decentralized
dynamic scheduling for grids and clouds using the community-aware scheduling algorithm. Future
Generation Computer Systems, 29(1), 402-415.
[49] Durillo, J. J., Nae, V., &Prodan, R. (2014). Multi-objective energy-efficient workflow scheduling
using list-based heuristics. Future Generation Computer Systems, 36, 221-236.
[50] Ghafarian, T., &Javadi, B. (2014). Cloud-aware data intensive workflow scheduling on volunteer
computing systems. Future Generation Computer Systems.
[51] Vasile, M. A., Pop, F., Tutueanu, R. I., Cristea, V., &Kołodziej, J. (2014). Resource-aware Hybrid
scheduling algorithm in heterogeneous distributed computing.Future GenerationComputer Systems.
AUTHORS
Jyoti Thaman is working as Assistant Professor in Department of Computer Science and Engineering, in M.
M. University Mullana, Ambala. She obtained her Masters from Kurukshetra University, Haryana, India, in
2007. Her interest lies in Cloud and Grid Computing. She is Ph. D Candidate from the Department of
Computer Science and Engineering, M. M. University, Ambala, India. She has published three research
papers.
Dr. Manpreet Singh is working as Professor and Principal, M M Group of Institutions, M. M. University,
Sadopur, Ambala, India. He obtained his doctorate from Kurukshetra University in 2010. He completed his
M. Tech and B. Tech.from Kurukshetra University, India. He has teaching and research experience of more
than 16 years. He has published more than 40 research papers in leading journals and conferences. His
research areas are Grid Computing, Cloud Computing and Ad hoc Networks.

Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review

More Related Content

Similar to Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review (20)

More from ijfcstjournal (20)

Recently uploaded (20)

Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review