“COMPARISON OF PARTICLE SWARM OPTIMIZATION AND GENETIC ALGORITHM FOR LOAD BALANCING IN CLOUD COMPUTING ENVIRONMENT”

(1)

“COMPARISON OF PARTICLE SWARM OPTIMIZATION AND GENETIC ALGORITHM FOR LOAD BALANCING IN CLOUD COMPUTING ENVIRONMENT”

K PATHAK, G VAHINDE

SSGMCE, Dept. of Information Technology, SHEGAON (MH)

[email protected], [email protected]

ABSTRACT: Cloud computing is a recent advancement where in IT infrastructure and applications are provided as ‘services’

to end-users. Cloud computing is the delivery of computing services over the internet. It is always required to share work load among the various nodes of the distributed system to improve the resource utilization and for better performance of the system.

So there may be load on VM’s, for balancing the load we have two ways, one is allocation of resources and second is scheduling the task. In this paper we are comparing two well-known non-linear algorithm for load balancing in cloud environment for effective utilization of the system and proper servicing the client’s request. We conducted our experiments on CloudSim simulator taking makespan as parameter for comparing the results of the different algorithm.

Keywords: Cloud computing, task scheduling, load balancing, PSO, Swarm optimization, Genetic algorithm, CloudSim, IaaS.

1. INTRODUCTION

Cloud computing is not another thought in specialized world but rather it is a forthcoming innovation. Grid Computing, Utility Computing and dispersed frameworks have direct association with the Cloud Computing. It can be expressed that framework processing goes about as the spine to Cloud Computing. Cloud computing gives virtual assets and administrations with the goal of diminishing expense. Cloud computing is actualized and famous generally because of its properties of giving virtualization and reflection [1] [2].

In operating system, various strategies have been projected for the purpose of job scheduling. The algorithms or strategies proposed for job scheduling are SJF (Shortest Job First), FIFO (First in First Out), LIFO (Last in First Out), Priority Based, Greedy Algorithm. The basic aim of these algorithms is minimization of total execution time of all jobs. These algorithms are easily understandable and can be implemented [3]. In Cloud environment, there is no restriction on number of jobs at a time requesting for scheduling which becomes the issue of efficiency of existing operating system scheduling algorithms. These algorithms may produce unwanted results in Cloud Computing environment and thus are not feasible to be implemented.

To make use of existing algorithms, there is a need of optimization of these algorithms to generate better results. One more issue in job scheduling and load balancing is that the response time for various task is very high and load on the processor become a threaten of failure of processor. This leads to a need of the algorithm which can optimize the load balancing process [4]. In this paper, both the Genetic Algorithm [5] and Particle Swarm Optimization approach [6]

has been implemented for scheduling and load balancing and a

comparison is drawn on the basis of defined parameters to find the better approach for scheduling in Cloud computing environment.

The paper’s organization is as follows:

• Section 2 describes the architecture of cloudsim.

• Section 3 describes the overview of the literature.

• Section 4 describes the comparison of PSO and GA for load balancing.

• Section 5 describes the conclusion and analysis.

2. INTRODUCTION TO CLOUDSIM

CloudSim goal is to provide a generalized and extensible simulation framework that enables modeling, simulation, and experimentation of emerging Cloud computing infrastructures and application services, allowing its users to focus on specific system design issues that they want to investigate, without getting concerned about the low level details related to Cloud- based infrastructures and services. Support for user-defined policies for allocation of hosts to virtual machines and policies for allocation of host resources to virtual machines

2.1 Architecture of CloudSim

CloudSim is implemented using existing simulation libraries and frameworks such as GridSim and SimJava in order to handle the low-level requirements of the system as shown in Figure 1. In particular, the components of handling events and message passing in SimJava can be reused in CloudSim.

Similarly, using GridSim simplifies the implementation of networking, information services, files, users, etc. We now discuss main components of a cloud computing infrastructure which are implemented in the simulator CloudSim. Figure below shows complete components of CloudSim.

(2)

Copy Right to GARPH Page 76 Figure 1. Layered CloudSim architecture

a. Data center: In CloudSim, data center is used to model the core services at the system level of a cloud infrastructure.

It consists of a set of hosts which manage a set of virtual machines whose tasks are to handle "low level"

processing, and at least one data center must be created to start the simulation.

b. Host: This component is used to assign processing capabilities (which is specified in the millions of instruction per second that the processor could perform), memory and a scheduling policy to allocate different processing cores to multiple virtual machines that is in the list of virtual machines managed by the host.

c. Virtual machine: This component manages the allocation of different virtual machines different hosts, so that processing cores can be scheduled (by the host) to virtual machines. This configuration depends on particular application, and the default policy of the allocation of virtual machines is "first-come, first-serve".

d. Datacenter broker: The responsibility of a broker is to meditate between users and service providers, depending on the requirement of quality of service that the user specifies. In other words, the broker will identify which service provider is suitable for the user based on the information it has from the Cloud Information Service, and negotiates with the providers about the resources that meet the requirement of the user. The user of CloudSim needs to extend this class in order to specify requirement in their experiments.

e. Cloudlet: This component represents the application service whose complexity is modeled in CloudSim in terms of the computational requirements.

f. Cloud Coordinator: These components manages the communication between other Cloud Coordinator services and brokers, and also monitor the internal state of a data center which will be done periodically in terms of the simulation time. All above are the entities related to the CloudSim which helps us to implement the Load balancing algorithm? Load balancing must take into account the two major tasks-one is the resource allocation and other is the task scheduling in distributed environment.

3. LITERATURE SURVEY

Radojevic and Zagar [4] proposed another calculation for burden adjusting called as CLBDM (Central Load Balancing Decision Module). The plan was proposed with the purpose of correspondence with all parts of PC structure, including workload balancers and application servers. CLBDM has effect sending choices on the heap balancers taking into account the collected data and inner estimations. It has determined that the execution of the composed structure can depend basically on isolating up work successfully over the taking an interest hubs in a circulated system of registering frameworks.

Ebehart and Kennedy [6] introduced particle swarm optimizer in an innovative structure. It has mentioned that genetic algorithm was much similar to particle swarm optimization.

Likely genetic algorithm, particle swarm optimization also begun with the generation of population but unlike GA, it can assume particles and can initiate them with some random position and velocity and allows them to move freely in search space. It has implemented the technique over various applications and concluded that PSO has given better

(3)

performance than other techniques. For example, due to crossover operator in GA, the immigration among subspecies of robots can be a serious issue; this issue should not be present with particle swarms.

AV. Karthick et al. [7] proposed a multi-queue scheduling scheme which can increase the client satisfaction and utilize energy of the system. Scheduling was the most important complex part in Cloud Computing and thus the major goal of global scheduler was to share the resources at most the maximum level. Researchers have given more importance to build a job scheduling algorithms that were well-suited and appropriate in Cloud Computing situation. Also in Cloud Computing, the user has to pay for services based on usage time that’s why Job scheduling was one of the critical event in Cloud Computing. Swachil J. Patel et al. [8] has raised an issue of priority in job scheduling because some job requests have need of being scheduled first then all other remaining jobs which can adjust with longer waiting time. With this aim, author has presented an improvement in job scheduling scheme based on the priority in Cloud Computing. Also this algorithm need to be further improved for the minimization of make span.

Daji Ergu et al. [9] proposed task-oriented resource allocation model in Cloud Computing environment. Due to different computers with varying capacities in Cloud environment, allocation of resources becomes complex and difficult. The proposed scheme has formed a pair matrix of tasks and comparison was made on the basis of network bandwidth, cost of task, its reliability and its completion time. The main motive was to improve the consistency ratio when allocation was based on weights of tasks.

Tingting Wang et al. [10] has discussed that the load balancing issue was critical in Cloud scenario due to huge number of users and large data volume. Thus the requests of resource sharing and reuse were becoming more and more imperative.

With the purpose of an efficient task scheduling strategy author has implemented load balancing using genetic algorithms so as to fulfill the user necessities and get better resource utilization. This strategy not only resulted into task scheduling sequence with shorter job and average job makespan, but has also satisfied the inter-nodes load balancing. But this strategy has assumed that there was no priority among jobs. However, in real Cloud Computing environment, it is not avoidable.

Ayed Salman et al. [11] has exhibited another undertaking task calculation that was in light of the standards of PSO in the circulated or parallel registering frameworks. PSO has taken after a populace based inquiry, which performs as indicated by

the easygoing conduct of winged animals and fishes. PSO framework has joined neighborhood seek techniques with worldwide hunt strategies. Through self-experience and neighboring knowledge, it endeavors to adjust investigation and abuse. Every individual component of the populace was called as a molecule that flies around in a multidimensional pursuit space in the hunt of the best arrangement. Particles may overhaul their position as per their own particular and their neighboring particles position, sending toward their best position or their neighbor's best position.

4. COMPARISON OF PSO AND GA FOR LOAD BALANCING.

4.1 Genetic Algorithm

A Genetic Algorithm is a search technique used in computing to find true or approximate solutions to optimization and search problems. Genetic Algorithms are categorized as global search heuristics. Genetic Algorithm are a particular class of evolutionary algorithms that use techniques inspired by evolutionary biology such as inheritance, mutation, selection, and crossover (also called recombination).

Genetic algorithm was the primary organic technique that supported the principle of survival and Mendel’s laws of inheritance. Genetic Algorithm has various benefits over alternative techniques for computationally intensive issues, if supplied with properly set operators and fitness functions.

Genetic Algorithm defines a collection of solutions (chromosomes) that are jointly referred to as a population.

The tactic then performs crossover, mutation and selection operations iteratively until the stopping criteria is satisfied [1].

The resultant set is the set of solutions.

Step 1: Selection

We randomly choose a set of the people based on their fitness:

Step 2: Crossover Step 3: Mutation

Algorithm for Genetic Algorithm

Step 1: Create a population of a fixed range of random chromosomes.

Step 2: Calculate the fitness values for all chromosomes.

Step 3: Choose 2 chromosomes having best fitness as parents (Selection step).

Step 4: Perform crossover among the parents to produce offspring using crossover ratio (Crossover step).

Step 5: Perform mutation if required at each position using mutation ratio

Step 6: Add the offspring chromosomes to the first population

(4)

Copy Right to GARPH Page 78 .

Step 7: If the termination condition is satisfied, then stop and come back to best chromosome as an answer, else repeat from Step two.

Flow Chart for Genetic Algorithm:

4.2 Bee Colony Optimization:

The BCO rule is predicated on the activities of bees whereas looking for nectar, and sharing the knowledge with alternative bees. There square measure 3 styles of agents - the employed bee, the onlooker bee, and the scout. The employed bee stays on a food supply and provides its surroundings in memory; the onlooker acquires this knowledge from the employed bees and selects one in every of the food sources to forage; and also the scout performs the task of finding new nectar sources.

The procedure for BCO is as follows:

Step1. Initialization: Distribution of the populations into the answer house randomly, and analysis of their fitness values, within the quantitative relation ne that represents the share of used bees within the total population.

Step2. Movement of Onlookers: Calculation of the likelihood of selection of a food source by the equation,

……… (1)

Where is the fitness value of the solution i evaluated by the employed bee, N represents the number of employed bees, and

is the probability of selecting the employed bee.

Choice of a target food source to move to onlooker bees and determination of the nectar amounts, by the equation.

……… (2)

Step3. Movement of Scouts: If the fitness values are not improved by continuous iterations, those food sources are abandoned, and these employed bees convert into scouts and are moved by the equation

Where r is a random number and r є [1, 2].

Step4. Updating of the Best Food Source: Memorization of the best fitness value and its position.

Step5.Termination Checking: Checking of the termination condition, if satisfied, termination of the program and output of the results; otherwise repetition of Step 2.

Evaluation of Filters Value Initialization of population

Apply Mutation Operator Selection of Parents

Apply Selection Operator

Offspring

Update Generation

Figure 2 Flow Chart for Genetic Algorithm

(5)

Table 1 Comparison of PSO algorithms for Load Balancing in Cloud Environment Ref.

wor k

Abstract Encod

ing schem e

Initial popula tion genera tion

Optimiz ation criteria

Natu re of tasks

Enviro nment

Highligh ts

Conclusion

[12] PSO algorithm embedded in crossover and mutation and in the local research, the experiment results show the PSO algorithm not only converges faster but also runs faster than the other two algorithms in a large scale.

1 ∗ n Vecto r Repre sentati on

Rando m

Commun ication Time and Executio n Cost

Indep ende nt

Cloud Simulati on Environ ment

– PSO algorithm both gains optimal solution and converges faster in large tasks. It is obvious that PSO is more suitable to cloud computing.

[13] A heuristic approach based on particle swarm optimization algorithm is adopted to solving task scheduling problem in grid environment. Each particle is represented a possible solution, and the position vector is transformed from the continuous variable to the discrete variable.

Rando m

Make span

Indep ende nt

Grid Simulati on Environ ment

Local search based on VNS applied after each permutati on

Simulation results demonstrate that PSO algorithm can get better effect for a large scale optimization problem.

Task scheduling algorithm based on PSO algorithm can be applied in the computational grid environment.

[14] User applications may incur large data retrieval and execution costs when they are scheduled taking into account only the ‘execution time’. In addition to optimizing execution time, the cost arising from data transfers between resources as well as execution costs must also be taken into account. We compare the cost savings when using PSO and existing ‘Best Resource Selection’ (BRS) algorithm. Our results show that PSO can achieve: a) as much as 3 times cost savings as compared to BRS, and b) good distribution of workload onto resources.

Rando m

Commun ication Cost and Executio n Cost

Work flow

Cloud Simulati on Environ ment (Swarm package )

Combine d with Heuristic

PSO based task-resource mapping can achieve at least three times cost savings as compared to BRS based mapping for our application workflow.

In addition, PSO balances the load on compute resources by distributing tasks to available resources.

[15] Here we propose a hybrid- scheduling algorithm to solve the independent task scheduling problem in grid computing. We have combined PSO with the gravitational emulation local search (GELS) algorithm to form a new method, PSO–

GELS.

Rando m while conside ring constra ints

Executio n Cost with Deadline Constrai nt

Work flow

Cloud Simulati on Environ ment (Java)

Used hill climbing after each iteration

The hybrid PSO algorithm performs better for local searches.

Because GELS is used for the local search rather than other local search algorithms such as hill- climbing or SA, the hybrid algorithm finds better solutions than other algorithms. PSO–GELS perform better than the other algorithms.

(6)

Copy Right to GARPH Page 80 [16] Here we present a Hybrid

Particle Swarm Optimization (HPSO) based scheduling heuristic to balance the load across the entire system while trying to minimize the makespan of a given task sets.

Rando m

Make span, No. of tasks that miss their deadline

Indep ende nt

Java Simulati on Environ ment

Local search based on GELS applied on results obtained from PSO

It is found that HPSO based task scheduling can achieve better load balancing as compared to PSO based scheduling.

[17] Here The representations of the position and velocity of the particles in conventional PSO is extended from the real vectors to fuzzy matrices. The proposed approach is to dynamically generate an optimal schedule so as to complete the tasks within a minimum period of time as well as utilizing the resources in an efficient way. We evaluate the performance of the proposed PSO algorithm with a Genetic Algorithm (GA) and Simulated Annealing (SA) approach.

Fuzzy Matric es

Rando m

Make span

Indep ende nt

Grid Simulati on Environ ment

Applying LJFP- SJFP heuristic alternativ ely after allocation of batch of jobs to nodes

We evaluated the performance of a fuzzy PSO for grid job scheduling and compared the performance with GA and SA. Empirical results revealed that the proposed approach can be applied for job scheduling. When compared to GA and SA, an important advantage of the PSO algorithm is its speed of convergence and the ability to obtain faster and feasible schedules.

Re f.

wo rk

Abstract

Enco ding sche me

Initial populat ion generat ion

Optimiz ation criteria

Selecti on operat or

Cros sove r oper ator

Mutat ion operat or

Nature of tasks and enviro nment

Conclusion

[18] This paper proposes a novel load balancing strategy using Genetic Algorithm (GA).

The algorithm thrives to balance the load of the cloud infrastructure while trying minimizing the make span of a given tasks set.

Fixe d bit strin g repre senta tion

Random ly

Load Balancin g

Rando mly

Sing le Poin t Cros sove r

Bits are toggle d (0â€“1 or 1â€“0)

Indepe ndent and Cloud Analyst

Analysis of the results, indicates that the proposed strategy for load balancing not only outperforms a few existing techniques but also guarantees the QoS requirement of customer jobs. However variation of the crossover and selection strategies could be applied as a future work for getting more efficient and tuned results.

[19] This paper presents a decentralized scheduling

algorithm based on genetic algorithms for the problem of DAG scheduling and the integration platform for the proposed algorithm in Grid system

Per muta tion Base d Repr esent ation

Random Make span, Load Balancin g, Resource Utilizatio n, Time taken to obtain solution

Roulett e Wheel

No oper ator

Adapti ve Mutati on Operat or

Workfl ow and Grid Simula tion

The DAG Scheduling algorithm has demonstrated to offer the best schedule length over all resources in a slightly shorter time than the other scheduling algorithms. The distribution of the tasks to the available resources has proved to accomplish good load balancing and efficient resource allocation by minimizing the idle times on the processing elements.

[20] This paper proposes a new scheduler which makes a scheduling decision

Dire ct

Resourc e having some data for

Makespa n

Roulett e Wheel

One- Poin t Cros

Flip Mutato r

Indepe ndent and Cloud

Simulation experiments show the effectiveness of GA-based load balancing technique method compared to FIFO and delay

(7)

by evaluating the entire group of tasks in the job queue. A genetic algorithm is designed as the optimization method for the new scheduler.

the task is assigned to task

sove r

Simula tion Enviro nment(

Hadoo p MapRe duce)

scheduling.

[21] To make the most efficient use of the resources, we

propose an

optimized scheduling

algorithm to achieve the optimization or sub-optimization for cloud scheduling problems.

Scheduling policy achieved by Parallel Genetic Algorithm which is much faster than traditional Genetic Algorithm.

Dire ct

Not mention ed

Resource Utilizatio n<comm a> Speed of resource allocatio n

Not mentio ned

Not ment ione d

Not mentio ned

Indepe ndent and Cloud Simula tion Enviro nment (using Java Langua ge with Java Geneti c Algorit hm Packag e)

Mathematically, we consider the scheduling problem as an Unbalance Assignment Problem.

Comparing to the GA, with parallel GA improved the speed of finding the best allocation.

And the utilization rate is better than the Round robin algorithm and Greedy algorithm.

[22] In this paper, we proposed a hybrid heuristic method (HSGA) to find a suitable scheduling for workflow graph, based on genetic algorithm in order to obtain the response quickly moreover optimizes makespan, load balancing on resources and speedup ratio.

Dire ct

New method Based on best- fit and round- robin method

Makespa n<comm a> Load Balancin

g on

Resource s<comma

>

Speedup Ratio

No operat or

Ran dom Gen e Sele ction Cros sove r

Selecti ng a rando m gene and replaci ng its resour ce with a resour ce having better failure rate and not overlo aded

Workfl ow and Cloud Simula tion Enviro nment

This Paper is based on GA, that uses the optimize characteristics Best-Fit and Round-Robin algorithms. For making the initial population, it uses two stages evaluation. At first stage, all tasks of application will be ordered by a priority method with respect to their influence to each other based on graph topology. In the second stage, the candidate resources will be assigned by combination of features of two methods, Best-Fit and Round robin to select good candidate resources.

[23] In order to solve this problem,

considering the new characteristics of cloud computing and original adaptive genetic algorithm (AGA), a new scheduling algorithm based on double-fitness adaptive algorithm- job spanning time and load balancing

Dire ct

Greedy algorith m

Makespa n<comm a> Load Balancin g

Rotatin g Selecti on Strateg y

One Poin t Cros sove r

Local Search

Indepe ndent and MATL AB

In this paper, we have proposed JLGA algorithm to achieve tasks scheduling with least makespan and load balancing. We have experimentally evaluated performance of JLGA algorithm and the DGA with 30 jobs. The experimental results shows that the JLGA takes little both in total job and average job consuming, and it balances the entire system load effectively.

(8)

Copy Right to GARPH Page 82 genetic algorithm

(JLGA) is

established.

[24] This paper presents a scheduling strategy on load balancing of VM resources based

on genetic

algorithm. This strategy computes ahead the influence it will have on the system after the deployment of the

needed VM

resources and then chooses the least- affective solution, through which it achieves the best load balancing and reduces or avoids dynamic migration.

Tree Repr esent ation

Spannin g tree based method

Load Balancin g

Rotatin g Selecti on strateg y

New ly deve lope d meth od

Newly develo ped metho d

Indepe ndent and Cloud Experi mental Enviro nment

The method achieves the best load balancing and reduces or avoids dynamic migration thus resolves the problem of load unbalancing and high migration cost caused by traditional scheduling algorithms. The experimental results show that this method can better realize load balancing and proper resource utilization

[25] Multi-agent genetic algorithm (MAGA) is a hybrid algorithm of GA, whose performance is far superior to that of the traditional GA.

This paper

demonstrates the advantage of

MAGA over

traditional GA, and then exploits multi- agent genetic algorithms to solve the load balancing problem in cloud computing

Fixe d bit strin g repre senta tion

Random ly

Load balancing

Neighb orhood Compe tition Operat or for agents

Neig hbor hood Orth ogon al Cros sove r Oper ator for agen ts

Mutati on operat or for agents

Indepe ndent and Cloud Simula tion Enviro nment.

This paper experimentally proves that MAGA is more appropriate than GA to handle high-dimensional function optimization problems. Then, establishing a cloud computing load balancing model, Min_min and MAGA algorithms were applied for resource scheduling respectively. This method, used for solving load balancing strategy based on virtualized cloud computing, is feasible and effective.

[26] In this paper, we present a hybrid approach called FUGE that is based on fuzzy theory and a genetic algorithm (GA) that aims to perform optimal load balancing considering

execution time and

cost. FUGE

algorithm assigns jobs to resources by considering virtual machine (VM) processing speed, VM memory, VM bandwidth, and the job lengths.

Dire ct

Random Makespa n<comm a>

Cost<co mma>

Load Balancin g

Based on fitness value

Fuzz y base d cross over appr oach

â€“ Indepe ndent and CloudS im, MATL AB

We compared the performance of our approach to several other cloud job scheduling models; the results of the experiments proved the efficiency of our FUGE approach in terms of execution time, execution cost, and average degree of imbalance. The FUGE approach modifies the SGA with the use of fuzzy theory and improves over the SGA performance in terms of execution cost by about 45% and in terms of total execution time by about 50%, which are the main goals of this research.

(9)

[27] This paper mainly focuses on how to improve the energy efficiency of servers in a data center by appropriate task scheduling

strategies. Based on MapReduce, Google’s massive data processing framework, a new energy-efficient task scheduling model is proposed in this paper. To solve this model, we put forward an effective genetic algorithm with practical encoding and decoding methods and specially designed genetic operators.

Dire ct

Random Energy Conserva tion

Rando m

Mult ipoin t Cros sove r

Single- Point Mutati on Operat or

VM Placem ent and Cloud Simula tion Enviro nment

We design a practical encoding and decoding method for the individuals, and construct an overall energy efficiency function of the servers as the fitness value of the individual.

Also, in order to accelerate the convergent speed and enhance the searching ability of our algorithm, a local search operator is introduced. Finally, the experiments show that the proposed algorithm is effective and efficient.

5. CONCLUSION

In this paper, we have surveyed GA and PSO method for load balancing in cloud computing. The main purpose of load balancing is to satisfy the customer requirement by distributing load dynamically among the nodes and to make maximum resource utilization by reassigning the total load to individual node. This ensures that every resource is distributed efficiently and evenly. So the performance of the system is increased. We have also discussed architecture of CloudSim simulator and required qualitative matrix for load balancing.

6. REFERENCES:

1. Dinh, Hoang T., et al. "A survey of mobile cloud computing:

architecture, applications, and approaches." Wireless communications and mobile computing 13.18 (2013):

1587-1611.

2. Zhang, Shufen, Hongcan Yan, and Xuebin Chen. "Research on key technologies of cloud computing." Physics Procedia 33 (2012): 1791-1797. Komal et al., International Journal of Advanced Research in Computer Science and Software Engineering 5(6), June- 2015, pp. 502-507 © 2015, IJARCSSE All Rights Reserved Page | 507

3. Silberschatz, Abraham, Peter Baer Galvin, and Greg Gagne

“Operating System Principles” John Wiley & Sons, 2006.

4. Radojevic, Branko, and Mario Zagar. "Analysis of issues with load balancing algorithms in hosted (cloud) environments." MIPRO, 2011 Proceedings of the 34th International Convention. IEEE, 2011.

5. Gu, Jianhua, et al. "A new resource scheduling strategy based on genetic algorithm in cloud computing environment." Journal of Computers 7.1 (2012): 42-52 6. Eberhart, Russ C., and James Kennedy. "A new optimizer

using particle swarm theory" Proceedings of the sixth

international symposium on micro machine and human science. Vol. 1. 1995.

7. AV.Karthick “An Efficient Multi Queue Job Scheduling for Cloud Computing”, International conf. on Computing and Communication Technologies (WCCCT), March 2014, 164 – 166.

8. Patel, Swachil J., and Upendra R. Bhoi. "Improved Priority Based Job Scheduling Algorithm in Cloud Computing Using Iterative Method" Advances in Computing and Communications (ICACC), 2014 Fourth International Conference on IEEE, 2014.

9. Ergu, Daji, et al. "The analytic hierarchy process: task scheduling and resource allocation in cloud computing environment" The Journal of Supercomputing 64.3 (2013):

835-848.

10. Wang, Tingting, et al. "Load Balancing Task Scheduling Based on Genetic Algorithm in Cloud Computing"

Dependable, Autonomic and Secure Computing (DASC), 2014 IEEE 12th International Conference on. IEEE, 2014.

11. Salman, Ayed, Imtiaz Ahmad, and Sabah Al-Madani.

"Particle swarm optimization for task assignment problem." Microprocessors and Microsystems 26.8 (2002): 363-371.

12. Lizheng Guo1et al., “Task Scheduling Optimization in Cloud Computing Based on Heuristic Algorithm”, journal of networks, vol. 7, 3, march 2012.

13. Lei Zhang1 et al., “A Task Scheduling Algorithm Based on PSO for Grid Computing”, International Journal of computational Intelligence Research. ISSN 0973-1873 Vol.4, No.1, 2008

14. Suraj Pandey et al., “A Particle Swarm Optimization-based Heuristic for Scheduling Workflow Applications in Cloud

(10)

Copy Right to GARPH Page 84 Computing Environments”, 24th IEEE Int. Conf. adv inf

netw appl, 2010.

15. Zhangjun Wu1 et al., “A Revised Discrete Particle Swarm Optimization for Cloud Workflow Scheduling”, IEEE transactions on cloud computing, vol. 2, April-June 2014.

16. Zahra Pooranian, “An efficient meta-heuristic algorithm for grid computing”, TELKOMNIKA Indones J Electr Eng 2012.

17. Gomathi B et al. “task scheduling algorithm based on hybrid particle swarm optimization in cloud computing environment” Journal of Theoretical and Applied Information Technology, 10th September 2013.

18. Kousik Dasgupta, Brototi Mandal et al., “A Genetic Algorithm (GA) based Load Balancing Strategy for Cloud Computing”, International Conference on Computational Intelligence: Modeling Techniques and Applications, (CIMTA) 2013.

20. Yujia Ge, Guiyi Wei, “GA-Based Task Scheduler for the Cloud Computing Systems”, International Conference on Web Information Systems and Mining, 2010.

21. Zhongni Zheng ,Rui Wang et al., “An Approach for Cloud Resource Scheduling Based on Parallel Genetic Algorithm”, 978-1-61284-840-2/11/$26.00 IEEE,2011.

22. Arash Ghorbannia Delavar, Yalda Aryan, “HSGA: a hybrid heuristic algorithm for workflow scheduling in cloud systems”, Springer Science+Business Media New York, 2013.

23. Tingting Wang, ZhaobinLiu et.al. “Load Balancing Task Scheduling based on Genetic Algorithm in Cloud Computing”, 12th International Conference on Dependable, Autonomic and Secure Computing IEEE, 2014.

24. Jianhua Gu, Jinhua Hu, Tianhai Zhao,Guofei Sun, “A New Resource Scheduling Strategy Based on Genetic Algorithm in Cloud Computing Environment”, JOURNAL OF COMPUTERS, VOL. 7, NO. 1, JANUARY 2012.

25. Kai Zhu, Huaguang Song et al., “Hybrid Genetic Algorithm for Cloud Computing Applications”, Asia - Pacific Services Computing Conference IEEE, 2011.

26. Mohammad Shojafar, Saeed Javanmardi et al., “FUGE: A joint meta-heuristic approach to cloud job scheduling algorithm using fuzzy theory and a genetic method”, Springer Science+Business Media New York, 2015.

27. Xiaoli Wang, Yuping Wang, Hai Zhu, “Energy-efficient task scheduling model based on MapReduce for cloud computing using genetic Algorithm”, JOURNAL OF COMPUTERS, VOL. 7, NO. 12, DECEMBER 2012.

7. AUTHOR PROFILE

K PATHAK Perceiving the 3rd year B.E. in Information And Technology at Shri Sant Gajanan Maharaj College of Engineering, Shegaon, Dist Bhuldhan, Maharashtra , India

G VAHINDE Perceiving the 3rd year B.E. in Information And Technology at Shri Sant Gajanan Maharaj College of Engineering, Shegaon, Dist Bhuldhan, Maharashtra , India