# Introduction he grid is the term defines based on the infrastructure of the distributed computing and offer the resources based on the client requirement. Grid technology can largely improve the virtual society's effectiveness and usefulness and supports the productivity. When going with improvement in grid technology, the grid facing many challenges by shared networking and collaboration and the main challenges by resource optimization and processes optimization. Grid computing technology is coordinated with the use of several numbers of servers; the grid server toil based on many applied techniques and methods. And these specialized gird servers in the network which acts together as a single logic integrated server system [1]. In 1969, Leonard Kleinrock was first visualized the concept of Grid in computing, he wrote: "We will probably see the spread of computer utilities, which, like present electric and telephone utilities, will service individual homes and offices across the country". In 1998, the redefinition of grid computing is evolved by Carl Kesselman and Ian Foster they wrote: "A computational grid is a hardware and software infrastructure that provides dependable, consistent, pervasive, and inexpensive access to high-end computational capabilities" [2]. The large-scale computation of the grid computing is the collection of heterogeneous autonomous system; systems are distributed throughout geographically, and the large number of heterogeneous networks is interconnected in the grid [3]. Fewer years back grid computing is formally defined as a technology that allows accessing, managing and strengthening the resources of IT in the environment of distributed computing. Grid computing is an advanced technology of distributed computing, that brings all databases, servers, infrastructures, applications and resources into a massive single system. The grid technology partnership among different enterprise organizations comprises the same organization and in addition external enterprise companies [1]. There are three reasons, which ensure grid computing is a promising technology [4]: The availability of the number of efficient computer resources which brings time-consuming and cost-effective to all clients, Grid computing can solve what normal system doesn't have the capability to solve some problem that can be solved by the available cooperative resources with massive computing power in the grid and Grid system directs the job to the proper resources of numerous computers can be run cooperatively and it works towards common goal with the usefulness of available resource collaboration which results the less time consumption and cost effective. In grid computing, there are lots of computers connected to grid to execute the jobs assigned by the clients, and among available computers at least a computer will perform as a server, this server takes all responsibility to allocate the client's jobs to the appropriate resource that are ready to execute [5]. Generally, grid resources can be divided into two types; one is software resources, and another one is hardware resources. In the category of the software resources which includes source of application pack, component services and data resources and the hardware resources which includes network resources, storage resources and computational resources [7]. The resources are distributed as grid geographically, unlike management policies are applied, heterogeneous # B Abstract -Grid computing, one of the most trendy phrase used in IT, is emerging vastly distributed computational paradigm. A computational grid provides a collaborative environment of the hefty number of resources capable to do high computing performance to reach the common goal. Grid computing can be called as super virtual computer, it ensemble large scale geographically distributed heterogeneous resources. Resource allocation is a key element in the grid computing and grid resource may leave at anytime from grid environment. Despite a number of benefits in grid computing, still resource allocation is a challenging task in the grid. This work investigates to maximize the profits by analyzing how the tasks are allocated to grid resources effectively according to quality of service parameter and gratifying user requisition. A fusion of SS-GA algorithm has introduced to answer the above raised question about the resource allocation problem based on grid user requisition. The swift uses genetic algorithms heuristic functions and makes an effective resource allocation process in grid environment. The result of proposed fusion of SS-GA algorithm ameliorates the grid resource allocation. resources are interconnected and all heterogeneous resources belong to various administrative domains. In grid computing, the term resource is defined as the capability that can be shared and utilized heterogeneous networked grid environment [6]. The pros of grid computing are ? Grid performance in processing data integration. ? Gird takes less time to solve more complex problems. In grid computing, large scale powerful resource allocation is a main challenge in a grid that may be critical to task performance. In general, the resource allocation in the grid faces many challenges in adaptability, load balancing, scalability and reliability [7]. The resources in the grid network are not controlled centrally and the resources can enter grid network and may leave anytime from the network autonomously. The autonomous property of the grid resources in the network leads to vagueness. The competence of the grid system is totally dependent on the proficiency of the resource allocation. As per the grid network, there is a substantial change in availability of the resource which varies the computational performance of the network and so there is a need for scheduling and allocation algorithm to survive from the changing environment for this network. To gear up the grid resource allocation, there is a need to overcome the challenges to speed up the processing power and resource memory in order to process the job in minimal computational cost as well as minimal computational time [8]. The First Come First Serve (FCFS) based resource allocation [19] that allocates the jobs which comes first. The other jobs in the job pool may wait longer due to the job size and resource availability for that particular size of job, and so this leads to very high time and cost consumption. The Shortest Job First (SJF) also named as Shortest Job Next (SJN) or Shortest Process Next (SPN) based resource allocation [19] that allocates the job which has shorter length on the fastest resource. When the continuous arrival of shorter length job pushes the job with the longest length leads to longer waiting time in the job pool. ? The Swift Scheduler (SS) based resource allocations [18] collects the job from the different users and swift the job using either by length or priority assigned. The resources are allocated based on the job swift by their priority in the job pool. The Genetic Algorithm (GA) based resource allocation is a population-based technique that allocates the jobs according to the fitness function evolved. The Genetic Algorithm (GA) performs crossover and mutation each in order to find the optimal solution. The Ant Colony Optimization (ACO) based resource allocation is a population-based technique which found the optimization problem solution using the pheromone values. Initially, the jobs are assigned to the resources randomly and each job verifies all the available resource using probability rule by assigning initial pheromone value and visibility value (i.e. simple heuristic value finds dividing one by distance between the job and the resource). After analyzing each job with each resource using probability rule, the process finds the reasonable solutions and stores values in the pheromone trails. The small amounts of pheromone trials are evaporated evaporation constant (?) and density of pheromone ( ). Finally the jobs are allocated to the resources which have the strong pheromone value and which found the optimal resource with less time consumption and low cost. The paper is structured as follows: The related works are declared in Section 2. The proposed fusion of Swift Scheduler and Genetic Algorithm (SS-GA) based resource allocation algorithm and its architecture are derived in Section 3. The experimental simulated resource allocation results are showed in Section 4. And Section 5 concludes the paper. # II. # Related Works A number of methods are being addressed by various researchers in the past; in this section some of the methods related to resource allocation are provided in this section. Fatos Xhafa et al. [3] have presented for the minimization of makespan and flowtime by designing the efficient Grid Scheduler with the usefulness of Genetic Algorithm (GA). Along with Genetic Algorithm two encoding schemes are used separately with GA, empirically studied and implemented. The pervious works are based only with the makespan estimation of the resource and this experimental study in the Grid Schedulers surpasses than the previous one by adding flowtime minimization with makespan minimization in a contemporaneous optimization mode. Furthermore, this encoding based Genetic Algorithms versions can able to discover the best and useful in real grids, which gives the best grid characteristics. This GA based grid scheduler can schedule the job arriving in the grid system dynamically and very fast by executing the work in the short time. Mayank Kumar Maheshwari et al. [9] have proposed the load balancing technique for proper distribution of the jobs. In general, Decision making, Information Collection and Data Migration are the three types of phases in Load Balancing. Load balancing algorithm which improves parallelism of work, proper distribution of task need to reduce the response time and resource utilization increases throughput management. Load balancing algorithm has two types of nature, static and dynamic. In this paper, they have proposed the optimal scheduling using Load balancing. This algorithm schedules the task by their minimum work completion time and to obtain load balance, it reschedules the waiting time of the works. This paper was based on the dynamic nature of the load balancing algorithm. By rescheduling the works according to the waiting time of the task tries to provide the best solution. It reduces the execution time of the job and with effective cost for the processing of all the jobs in the grid system. Karthick Kumar [10] This algorithm schedules the task by their fair completion time and to obtain load balance, it reschedules using mean waiting time of the works. This fair scheduling algorithm scheme tries to reduce the execution time and minimization of cost for all jobs provides best solution in the computational grid. Adjusted Fair Task Order, Earliest Deadline First, Max Min Fair Scheduling and Simple Fair Task Order algorithms are compared with the proposed algorithm by using simulation. Murugesan et al. [11] have proposed a resource allocation framework in a grid computing environment for heterogeneous workloads, subject to a set of conditions. The resource allocation approach can manage and assign the task to the available grid resources with the minimization of cost from user side. They proposed a mathematical model to allocate the user task in the job pool to the accurate resources in the grid resources pool with the purpose to reducing the grid user's costs with reference to the time limit and budget denoted by the grid user. In this paper, they presented a reasonable model to consider the Quality of Service (QoS) parameter requirements to complete the user task, and at the same time they examine the performance of this proposed algorithm. Manavalasundaram et al. [12] have proposed Grid association states the method for synchronized sharing of distributed clusters based on the computational economy, permits the grid users to use the local resources from the grids association but doesn't satisfies the users requirement. The computational economy methodology used here as organizing the resource which not only satisfying the Quality of Service (QoS) parameters, but also the best performance of the resources. The entire selfprovisioning for the each user's as autonomous world in Grid Association. The proposed efficient methodology for resource management can use by grid user effectively. The local resources in the grid association doesn't met the requirement of the grid user, so it the job migrate remote resources according to the user's condition in Quality of Service (QoS) requirements. The global scheduler manages and schedules the user job in the Virtual Organization (VO) and that imposes Virtual Organization (VO) -wide policies. The agents used in the grid association to access and maintain the shared directory of the association of the grid resources. The experimental results of this work shows that the resource which have high capability to execute the job in low mean time and cost effective, and so there is a very huge competitions from all other jobs in the association to that particular resource to satisfy their Quality of Service (QoS) requirements, the resource association provides a increased ability to satisfy Quality of service requirement, in general the resource allocation methodology provides an increased ability to satisfy Quality of service (QoS) parameter needs overall the grid users. Navjeet Kaur et al. [13] have proposed managing the resource using inter-intra fairness scheduler in grid environment. In grid computing environment an essential role is resource allocation management. The grid system responsibility is to make sure weather the client's jobs/applications request are getting the best resources in cost effective and timely manner. In grid computing technology, there are many resource allocation algorithm are there making jobs allocate to the resources allocation decisions. In this paper, they describes the resource management with different proportional share scheduler with O(1) precision in grid environment. While allocating the resource to the job, the resource allocation's fairness and efficiency that ensure by proportional/fair share scheduler. The proposed inter -intra fairness scheduler is the integration efficient fair share scheduler features as well as managing resources and job scheduling issues in a soft way. Rajkumar Buyya et al. [14] have identified challenges in grid resource management in the grid computing environment and proposed an effective resource allocation management and job scheduling as metaphor in the computational grid environment. Their projected algorithm analyzes the challenges and requirements in the distributed resource allocation of economy based grid system. The various agent economy based system considering both promising and historical. The CPU cycles, storage, and network bandwidth of the resources are considered by the various economy based system. It presented an extensible and leverage of existing resource management technology and service oriented grid architecture in the grid system. And it also presented auction models and merchandise for the resource management in the grid technologies. The use job scheduling and merchandise economy model for resource management in both data grids and operational is also presented. Zne Jung Lee et al. [15] have proposed cost effective resource allocation process by allocating the job to optimal resources in the grid environment. In this paper, they proposed the hybrid search algorithm heuristics approach for the resource allocation problem is encountered in grid system. The proposed algorithm can explore the search space and exploit the best solution with the dual advantages of Genetic Algorithm (GA) and Ant Colony Optimization (ACO) of population algorithms. The well designed GA and ACO are implemented for resource allocation problem. In addition, heuristics in Ant Colony Optimization are used Devaki et al. [16] have proposed population based Genetic Algorithm considering Quality of Service parameters for Job scheduling. The utilization of an idle resource and distributed resources present global grid system to solve the challenges which are computational was greatly encouraged in grid computing. The problems are either in two sides one in scheduling the job or another was executing jobs in available resource in the grid system. Here the main issue was scheduling the jobs in correct resource is considered as NP complete problem, and so it was essential to have an efficient job scheduling to be effective utilization of the resource in the grid system. Various heuristic algorithms are used to solve the issues which bring a nearby optimal solution. In this paper, they proposed an offline mode Genetic Algorithm evolution using Quality of Service (QoS) parameter satisfaction for scheduling the job to heterogeneous resources. The proposed algorithm mainly focused in makespan and Quality of Service (QoS) satisfaction, when selecting the optimal resource. Prabhu et al. [17] have proposed the resources scheduling approach based on the multi-objective Genetic Algorithm (GA). This proposed multi-objective Genetic Algorithm (GA) focused on the mean waiting time, flow time and optimal resource allocation and batch mode methods used for the allocation of jobs to the favorable resource. They evaluate the performance using the methods with multi-objective parameters based on their computational results. The overall execution time is minimized by Genetic Algorithm (GA) based scheduling. This proposed algorithm was tested with up to 1000 generation, and the experimental performance shows the results achieved near optimal efficiency. # III. The Proposed Algorithm a) The basics of Swift Scheduler and Genetic Algorithm The swift algorithm and genetic algorithm is the key algorithm in problem solving methods. The swift focus and works according the users requirement and the genetic algorithm is a population based search algorithm which found the optimal solution to the problem. Meta-Heuristic Search Function # i. Swift Scheduler # Basic Structure The Swift Scheduler based Resource Allocation works as follows: Begin Step 1: Initialize the N jobs in job pool; Step 2: Initialize the available M resource in Grid system; Step 3: Schedules the N jobs in ascending order based on the job's length; Step 4: Search the resource using meta-heuristic function based on less computational time; Step 5: Allocate the jobs to the appropriate resources; End The Genetic Algorithm based Resource Allocation defines the resource allocation process by analyzing the compatibilities between jobs and available resources in the Grid system. The compatibility was analyzed by the job's computational time and if the expected computational time was satisfied, the jobs will be allocated to the available grid resources. Fig. 3. shows the processing of genetic algorithm based resource allocation. The genetic algorithm can search the available resources in polynomial time by its metaheuristic search technique. The algorithm initializes the population by random selection of chromosomes. It can explore the appropriate resource by calculating fitness function for each chromosome and selecting the chromosomes for mating for next generation. Crossover was done between randomly selected chromosomes with fixed probability rate, and mutation was also done in the random manner with some fixed probability rate. These operations are repeated every iteration while it satisfies the job with expected computational time in available resources. The main objective of this resource allocation model is to allocate the proper jobs to the fittest resources in order to meet the Quality of Service (QoS) parameters like minimum job completion time, cost efficient, economy, and resource utilization. Our proposed grid resource allocation model combines the swift scheduling mechanism and genetic algorithm to make the resource allocation process as more proficient when compare with other methods. In our methodology, we take the input data set which contains jobs and resources to be processed. A job pool consists of the number of jobs having its id, length of the job and job priority while the resource pool has resource id, its capacity and cost to execute the particular job. The following assumptions can be made to achieve the resource allocation as the finest one. # Available Resource # Create Initial Population ? Every single job is autonomous of each other and there is a priority among them. ? Every resource and jobs is simultaneously available at the early stage of time. ? A resource can only process one job at a time and this procedure cannot be interrupted before completed. ? For every resource and job has its unique id to differentiate each other and avoid conflicts while processing jobs in the resources. ? The migration of jobs is not permitted. ? In job and resources, we need to specify the job length, job priority, resource capacity and resource cost to allocate effectively. In the proposed SS-GA algorithm, it fuses the swift scheduler with the population based Genetic Algorithm. So it makes full use of advantages of the priority based job scheduling offered by swift scheduler, as well as the powerful accurateness finder ability provided by the Genetic Algorithm. # B proposed algorithm accepts the dynamic nature of grid resource availability and users' task. # Begin Step 1: Initialize the jobs in the job pool; Step 2: Initialize the available resources in resource pool; Step 3: Schedule the job with high priority in ascending order; Step Step 2: Swift schedules the jobs in the job pool according to the user's importance i.e. priority; Step 2.1: Priority of the jobs are randomly assigned between 0 to 5, 0 holds the high priority of the jobs and some jobs doesn't hold priority; Step 2.2: Separate the jobs with ( ) and without ( ) priority in the job pool; Step 2. Display the results of total best time and cost of the allocated job; And find the resource utilization percentage and economy rate; Finally, compare all the results with the existing resource allocation algorithm. END iv. Swift Scheduler Process The fusion of swift scheduler and genetic algorithm (SS-GA) based resource allocation defines the resource allocation process in the fast resource scheduling manner. The word swift describes the capability of the work process to be done in high speed. The resource allocation algorithm uses the advantage of the swift algorithm by allocating the jobs in available resource according to user requirement. So the job with priority gives more importance for the resource allocation. The job assigned with priority denotes the urgency of job and '0' is considered as the highest priority among the jobs in job pool. The swift algorithm uses search algorithm to find the optimized available resource in the grid system. The search algorithm finds optimized available grid resource for the high-priority job; it makes an effective resource allocation. Here genetic algorithm used along with swift algorithm to find the optimized available resource in proposed grid system. Swift scheduler separates to the prioritized jobs, and non prioritized job in the job pool. And it arranges/schedules the prioritized job (i.e. priority numbered '0' is considered as high priority job) and the non-prioritized jobs are schedules the jobs according to its length in ascending order (i.e. Shortest Job First (SJF) order) as shown in the fig. 4. The jobs are arranged/scheduled using the following formula. (1) Where, denotes the job with priority, denotes the non-priority job, P denotes the priority of the job and l denotes the job's length. The resource allocation takes the full advantage of swift scheduler for fast execution of high prioritized job first. Swift scheduler increases the scalability of memory and computational time. So the swift algorithm ameliorates the resource allocation process. # v. Genetic Algorithm Process After performing the swift scheduling process, job tasks have scheduled according to its priority, and the procedure can be merged with genetic algorithm to achieve the objective function of less job completion time, cost-effective and its economy, and resource utilization. A Genetic Algorithm (GA) has four major steps: fitness, selection, crossover, and mutation. Genetic algorithm is capable of solve optimization problems by carrying out the genetic process. A possible solution to a definite problem may be represented as a chromosome containing a sequence of genes. Initially, the population size is a set of chromosomes is generated, and it undergoes many genetic processes. By using selection, crossover and mutation operations, GA is capable of progress towards the population to generate a best possible solution of resource allocation in grid environment. # a. Initial Chromosome Representation The GA operates on a population of chromosomes, which is encoded according to the problem. The chromosome represents a complete solution to the problem. The chromosome is a collection of n number of jobs to be initialized randomly with resources from 1 to where L = {1,2,3,??N}. Therefore, the number of chromosome sequence can be generated based on population size P(s). In this equation, is the length of the job and is the resource capacity to finish the certain job. From this, the overall job completion time ( ) can be calculated as it is the sum of execution of all the n jobs running on their corresponding allocated resources as, (2) (3) (4) (5) Therefore, the total cost occurred to complete the grid resource allocation can be derived as, the major role in well-organized allocation of resources because it only picks the best fitness chromosomes to undergo the genetic operations namely crossover and mutation. It selects the chromosomes having probability of best fitness values. If there are two or more chromosomes having the same best fitness, one of them is chosen randomly. Finally, there are n/2 chromosomes are selected for a genetic operation where n is the population size to produce new chromosomes. # d. Crossover Operation Crossover is a genetic operator that combines or mates' two chromosomes (parents) derived from selection process to generate a new child chromosome (offspring). The two chromosomes (strings) take part in the crossover is known as parent and the ensuing chromosomes are known as child strings. This process takes place by fixing the crossover rate as 60% to 70% for making this process a more effective. There are several types of crossover is carried out in genetic algorithm like one-point crossover, two-point crossover, arithmetic crossover, uniform crossover and heuristic crossover. Here the crossover function used is one-point crossover, because one-point crossover is suitable for task ordering problems. # One-Point Crossover This crossover function selects a particular point randomly in the parent chromosomes to be induced, and then it interchanges the two-parent chromosomes at this point to create two new offspring. A point which selected randomly plays the significant role in one-point crossover. This process carried entirely depends on the crossover rate . Let us consider that the two-parent chromosomes have been selected for undergone the crossover process as shown below. The crossover point of both parent chromosomes is denoted by varying the colors in the chromosome. The offspring (child) chromosomes can be produced by interchanging their strings according to the crossover point as, Figure 10 : One-Point Crossover e. If the chromosomes arrived after many steps of crossover, there is a chance of strings (chromosomes) having repeated or same one. In order to prevent the particular environment, the mutation is done to gather the best individuals among the parent chromosomes. The mutation has to be performed with the help of mutation rate . This could make creating the completely new genes among the chromosomes in population by altering (changing) gene values. Flip Bit is one of the most common types of mutation used as it changes the bit values of genes from 0 to 1 or vice versa. This mechanism can be illustrated as follows. The child can be produced by invert the gene values present in the above parent chromosomes. The pink color denotes the changed values which are done by Flip bit mutation. IV. # Experimental Results and Discussion In this section, the experimental results are discussed in simulated grid network using dotNet framework platform. The proposed algorithm is compared and evaluated with existing algorithm using certain parameters like resource utilization, computational time, processing time, makespan, economy, time, cost etc. And these results are evaluated using separate line graphs. The results are taken in dynamic nature of grid environment. The dataset are simulated which contains job table and resource table separately. The job table contains its id, Mutation length and cost, similarly the resource table also contains its id, capacity and its cost. # a) Resource Utilization The total resources consumed by the jobs against the expected amount of resources for a particular work. The utilization is normally calculated in percentage of time. Resource utilization is generally, the resource capability of the job execution in a period of time. Resource utilization will differ according to the resource processing speed. The idle resource should have very low utilization rate. The proposed algorithm boosts up all the resource in active stage in dynamic manner, the resources are utilized by the incoming tasks. The utilization formula is defined in equation 6. # b) Economy Economy is the word in grid resource allocation defines vigilant management and proficient or restrained purports the proficiency of larger resource allocation process with lower price. It will track the changes on the price level with the efficient use of resource price and can analyze and improve the performance of macroeconomic. The economy formula is defined in equation 7. # c) Processing Time The processing time of the tasks are calculated by the following equation 8. It checks the first job with allocated resource and took its length of the job that which execute in the resources. And the upcoming jobs are checked with the all resources when the resource id and resource id are equal, this has to be done similar manner to all jobs allocated in the resources and so on. The resource in the grid environment may connect and leave at anytime. So the results are taken in a dynamic way by varying job count and resource count in the simulated grid environment using dotNet framework. The proposed SS-GA algorithm with existing algorithms using utilization, economy, total resource cost, processing time and computational parameters are evaluated. The following results are compared between First Come First Serve These algorithms are compared using parameters like resource utilization, resource economy, total resource cost, processing time and computational time. The results are taken with different available resource pool size and in all cases of available resource the dynamic job inputs are compared in following table with proposed algorithm. The above five tables are compared with constant 10 resources with dynamic job inputs. The result with different job inputs explains that the resource utilization is comes best compared to other algorithms, the resources are not fully occupied and it is used effectively. The proposed algorithm maintains the resources and avoiding the resources in idle stage. The economy rate using the resource cost shows proficient resource usage. The most important and expecting parameter should be the cost, the performance of the proposed algorithm bring the effective cost while comparing other existing algorithm. The processing time and computational time are totally depends implemented code. # B The above five tables are compared with constant 15 resources with dynamic job inputs. When the number of resources goes high, the rate of resource utilization maintains its proficiency with the use of proposed algorithm. The resource utilization and economy shows the good rate while using the proposed algorithm. The fusion of SS-GA algorithm gives the computational time is pretty high because the fused algorithm takes both the work of swift algorithm and genetic algorithm. The above five tables are compared with constant 20 resources with dynamic job inputs. The economy results of SS-GA algorithm bring good rate, when increase in the number of available grid resources in the grid network. The total resource cost of proposed algorithm with varying resources results the effective cost. The performances of the resource allocation in the grid environment are boosted up by SS-GA algorithm. The processing time and computational time of the ACO based resource allocation algorithm is very low because it allocates the jobs by checking with the resources in first time itself. But the proposed algorithm takes the advantage of swift algorithm; it checks the job with priority brings higher order for allocation process. And so it takes more time for resource allocation computation. The available resources are dynamically checked by the proposed algorithm. The above five tables are compared with constant 25 resources with dynamic job inputs. In the table.4a shows the slight change in economy rate of proposed algorithm with AWPF and AWOPF and resource cost also ACO based resource allocation algorithm's very nearest to the proposed one. When there is increase in number of job in resource allocation process, the resource utilization and economy rate are being fine. And the small scheduling based algorithm i.e. FCFS, SFJ, SS are maintaining good results but particularly in some cases because it only gives big concentration in scheduling the jobs to resources and it won't consider the cost of the resources while resource allocation. The above five tables are compared with constant 30 resources with dynamic job inputs. The proficient resource allocation in grid computing and the raise in the performance show best result with the proposed fusion of SS-GA based resource allocation. And total resource cost while resource allocation process. V. # Conclusion In this paper, a mechanism was proposed to allocate the jobs efficiently to the corresponding resources during the process. The proposed method was implemented in the Microsoft Visual Studio 9.0 environment with dotNet framework. This paper has taken GA-SS based algorithm to achieve the quality of service in the resource allocation. In the first phase, the jobs in the job pool are to be scheduled according to its priority. The scheduling of these jobs was done by Swift scheduler. In the second phase, the scheduled jobs can obtain their corresponding resource to allocate by Genetic algorithm. The genetic algorithm performs major operations like fitness, crossover and mutation in the scheduled jobs to achieve less flow time and cost. The experimental result shows that the proposed GA-SS based scheme was much better than the other traditional methods like First Come First Serve (FCFS), Shortest Job First (SJF), Swift Scheduler (SS) and Ant-Colony Optimization (ACO) algorithms in terms of makespan, cost, resource utilization and economy. And also, the experimental results were verified that the proposed methodology can be adaptable to give good results while varying the number of jobs and resources in the input data set. ![Process the huge date sets into smaller one for faster execution. ? Put to good use the available heterogeneous hardware in the grid. ? Collaboration between the different organizations becomes easier. ? The cons of grid computing are ? Grid standards and software are still evolving one. ? The job submissions in grid network are noninteractive. ? Grid resource administrations are not properly controlled. ? No proper allocation of jobs to the appropriate resources. ? Due to the improper resource allocation leads to long execution time and added cost.](image-2.png "") ![In this paper, the above all algorithms are compared with proposed fusion of Swift Scheduler and Genetic Algorithm (SS-GA) based resource allocation as shown in fig.1. The Swift Scheduler and Genetic Algorithm (SS-GA) based resource allocation algorithm fuses prioritized jobs in the job pool using Swift Scheduler (SS), with the Genetic Algorithm (GA) based resource allocation, so it makes full use of advantages of the Swift Scheduler, estimated with a Genetic Algorithm, as well as ability provided the effective resource allocation. The swift scheduler arranges/schedules the jobs by the user's urgency (i.e. assigned as priority) or length of the job. And the arranged/scheduled jobs are allocated to the appropriate resources using Genetic Algorithm to obtain time-consuming and cost-effective as best. The Quality of Service (QoS) [14] parameter includes job execution time and cost efficiency are compared in First Come First Serve (FCFS) based resource allocation, Shortest Job First (SJF) based resource allocation, Swift Scheduler (SS) based resource allocation, Ant Colony Optimization (ACO) based resource allocation and AWOPF & AWPF Constraint checked Genetic Algorithm (GA) based Resource Allocation with the proposed fusion of Swift Scheduler and Genetic Algorithm (SS-GA) based resource allocation algorithm. The performance of the proposed fusion of Swift Scheduler and Genetic Algorithm (SS-GA) based resource allocation algorithm compared with the above mentioned other algorithms by using simulation.](image-3.png "") 1![Figure 1 : The comparative model of various algorithms with proposed fusion of Swift Scheduler and Genetic Algorithm (SS-GA) based Resource Allocation](image-4.png "Figure 1 :") ![Computational Profit in Grid Resource Allocation using Dynamic Algorithm Global Journal of Computer Science and Technology Volume XIII Issue II Version I to amend the search performance of the resource allocation problem. While testing this proposed algorithm by simulation, the results showed attractive performance for resource allocation problem.](image-5.png "MaximizingB") 2![Figure 2 : Swift Scheduling based Resource Allocation Swift scheduler based resource allocation defines the resource allocation process in the fast resource scheduling manner. Swift scheduler based resource allocation is the combination of the shortest job first and local meta-heuristic search function. Fig.2.shows the processing of swift scheduler based resource allocation. Swift scheduler gives the importance to the jobs, which schedules the jobs according to its length in ascending order (i.e. Shortest Job First (SJF) order). The resource allocation takes the advantage of metaheuristic search function. The heuristic function searches the resource which has the capability to execute the job in minimal job computational time.](image-6.png "Figure 2 :") 34![Figure 3 : Swift Scheduler based Resource Allocation Structure](image-7.png "Figure 3 :Figure 4 :") 5![Figure 5 : GA based Resource Allocation Structure b) Fusion of Swift Scheduler and Genetic Algorithm (Ss-Ga) Based Resource Allocation](image-8.png "Figure 5 :") 6![Figure 6 : Proposed SS-GA algorithm for Resource Allocation](image-9.png "Figure 6 :") 4![Fig. 7 illustrates the algorithmic structure of proposed fusion of Swift Scheduler and Genetic Algorithm (SS -GA) based resource allocation. The](image-10.png "Fig. 4 .") 4![If the job with no priority, schedules the job in shortest job first order; Step 5: Initialize the scheduled Population size P(s) based on the number of jobs; Step 6: Perform a Fitness function to minimize the following objective function; Step 6.1: Job Completion Time ( ); Step 6.2: Job Completion Cost ( ); Step 7: Selecting half of the individuals among the population P(s) having best fitness values; Step 8: Apply One-Point Crossover for the selected individuals with a crossover rate ; Step 9: Mutation to be carried out at the rate of to obtain the new better chromosomes; Step 10: Terminate the above process after 'n' iterations to achieve best results in case of objective function in resource allocation; End](image-11.png "4 :") 78![Figure 7 : Proposed SS-GA Algorithm Structure](image-12.png "Figure 7 :Figure 8 :") 3![With prioity jobs ( ): If P =! Null; For each P[j] < P[Min]; Min = j; ? If P[i] == P[j]; Consider the length of the jobs For each L[j] < L[Min]; Min = j; ? Arrange the jobs according to the priority of the job; Ptmp = P[i]; P[i] = P[j]; P[j] = Ptmp; Step 2.4: Without priority jobs ( ): If P == Null For each L[i] < L[Min]; Min = j; according to the length of the job; Ltmp = L[i]; L[i] = L[j]; L[j] = Ltmp; Generally, Step 2.5: Finally the chromosomes are arranged in the job pool by prioritized jobs followed by the without prioritized jobs; Step 3: Initialize the chromosomes scheduled by the swift scheduler; Step 3.1: Evaluation of fitness function Step 3.1.1: Time fitness For each job's completion time: Overall completion time: Step 3.1.2: Cost fitness For each Job's cost: Overall cost: Step 3.2: Best chromosome selection Here, c1 & cL are first chromosome and last chromosome respectively; For each selection & total chromosome length ; Here, cN represents the intial best chromosomes; Step 3.3: Chromosome Crossover , , Global Journal of Computer Science and Technology Volume XIII Issue II Version I](image-13.png "3 :") 9![Figure 9 : Chromosome Representation of Jobs and Resources based on Population size b. Evaluation of Fitness Function In genetic algorithm, fitness function is the most important step for optimize the solutions in grid resource allocation. Fitness is initially resulted from the objective function and used in consecutive genetic operations. The objective function may be minimized or maximized based on the problem occurs in various domains. The purpose of a fitness function is to afford a significant, quantifiable, and equivalent value given a set of genes present in the chromosome. Here the objective function is cost efficient and minimum job completion time. Fitness can be defined as the value assigned to an individual based on how far or close an individual is from the solution; greater the fitness value superior the solution it contains. It mainly considers two parameters those have to be optimized as follows Job Completion Time ( ) It can be defined as the time taken by a job to be executed successfully in the given allocated resource,](image-14.png "Figure 9 :=") 11![Figure 11 : Flip Bit Mutation](image-15.png "Figure 11 :") ![(FCFS) based Resource Allocation, Shortest Job First (SJF) based Resource Allocation, Swift Scheduler (SS) based Resource Allocation, Ant Colony Optimization based Resource Allocation, AWPF Constraint checked Genetic Algorithm (GA) based Resource Allocation, AWOPF Constraint checked Genetic Algorithm (GA) based Resource Allocation with the proposed fusion of Swift Scheduler and Genetic Algorithm (SS-GA) based Resource Allocation.](image-16.png "") 121317![Figure 12 : Comparison on Utilization with constant Job 50](image-17.png "Figure 12 :BFigure 13 :Figure 17 :") ![Rule Scheduling Algorithms Using Different Inter Arrival Time Jobs in Grid Environment", International Journal of Grid and Distributed Computing, Vol. 4, No. 3, 2011.](image-18.png "") .FCFSSJFSSACOAWPFAWOPFSS-GAUtilization101.13731101.13731101.1373196.8543104.0636100.329594.96753Economy35.9724135.9724135.9724139.244.241844.1238.86897Total Resource Cost2321232123212204204821242116Processing Time7412741274121502163015342163Computational Time63.495641.057615.37297.2747217.9281487.5417671.9201 10132Year22D D D D ) B(.a : Resource -10 Job -40 Maximizing Computational Profit in Grid Resource Allocation using Dynamic Algorithm Global Journal of Computer Science and Technology Volume XIII Issue II Version I © 2013 Global Journals Inc. (US) Maximizing Computational Profit in Grid Resource Allocation using Dynamic AlgorithmFCFSSJFSSACOAWPFAWOPFSS-GAUtilization115.0763115.0763115.076399.17764108.6022104.865898.36868Economy43.7648943.7648943.7648941.540943.690643.885641.40852Total Resource Cost2224222422242208213321982203Processing Time5399539953991152116212081242013 2Computational Time26.508340.063424.36623.3151230.8437498.984671.4895YearFCFSTable 2.a : Resource -15 Job -40 SJF SS ACOAWPFAWOPFSS-GA24Utilization93.1788993.1788993.17889100102.8329106.801795.01312Volume XIII Issue II Version IEconomy Total Resource Cost Processing Time Computational Time Utilization Economy37.70749 2777 10018 32.1723 FCFS 99.10011 39.7066937.70749 2777 10018 28.8015 Table 2.b : Resource -15 Job -50 37.70749 38.89637 2777 2724 10018 1664 28.069 4.2309 SJF SS ACO 104.3839 99.10011 102.9206 40.53162 39.70669 40.3116436.957 2685 1620 272.0314 AWPF 102.8136 40.145537.3555 2710 1633 543.2246 AWOPF 104.2056 40.621138.04397 2682 2056 738.77 SS-GA 97.45575 39.43172Global Journal of Computer Science and Technology ( D D D D )Total Resource Processing Time Computational Time Utilization Economy Total Resource Cost Processing Time Computational Time Cost16148 27.5003 FCFS 123.5437 42.60374 3824 22756 29.8284 FCFS 328915478 31.5784 Table 2.c : Resource -15 Job -60 16148 2203 22.5031 4.0919 SJF SS ACO 115.8134 115.8134 99.12366 41.73266 41.73266 39.38866 3879 3879 3827 23503 23503 2975 29.1116 45.4337 5.0326 Table 2.d : Resource -15 Job -70 SJF SS ACO 3264 3289 32562290 342.8629 AWPF 93.3148 39.8867 3819 2911 385.6059 AWPF 32922385 632.962 AWOPF 93.0657 39.7923 3884 2949 658.5975 AWOPF 33272720 787.1048 SS-GA 97.32314 39.08774 3811 3667 809.2128 SS-GA 3254Utilization95.70403111.808895.70403100.2521100.3407101.274499.58263Economy38.2131342.5861238.2131340.9311138.682738.878838.62347Total Resource Cost4592426745924390432943724338Processing Time3376532349337653900384239504436Computational Time35.703534.534633.27475.9805467.8766868.2343876.4405Table 2.e : Resource -15 Job -80© 2013 Global Journals Inc. (US) .FCFSSJFSSACOAWPFAWOPFSS-GAUtilization90.802890.802890.8028110.741498.0504100.327695.5254Economy37.8052237.8052237.8052241.0099939.030239.753939.28857Total Resource Cost4483448344834352433843574304Processing Time Utilization Computational TimeFCFS 28390 117.2917 38.908SJF 28390 117.2917 44.2368SS 28390 117.2917 40.7869ACO 2914 100.5357 8.2141AWPF 3079 87.6387 502.8877AWOPF 3153 86.9029 908.77813593 SS-GA 93.05785 856.5241Economy39.0922439.09224 Table 3.e : Resource -20 Job -80 39.09224 36.7496337.345636.876235.43192Total Resource Cost2180218021802160218122392155YearProcessing Time3532353235328648718961256Computational Time40.676366.905435.31754.2761241.6037514.3924705.31925Utilization Economy Total Resource Cost Processing Time Computational TimeFCFS 111.026 40.80768 2753 7349 40.7577Table 3.a : Resource -20 Job -40 SJF SS ACO 111.026 111.026 98.10555 40.80768 40.80768 38.88889 2753 2753 2739 7349 7349 1338 32.6365 54.1946 6.7784AWPF 103.9589 40.2273 2682 1354 309.8983AWOPF 99.7241 39.5456 2767 1321 595.1578SS-GA 93.06804 37.99643 2679 1808 732.8609Volume XIII Issue II Version ITable 3.b : Resource -20 Job -50D D D D D D D D )(FCFSSJFSSACOAWPFAWOPFSS-GAUtilization100.4372100.4372100.437297.24868107.554112.009894.25641Economy42.7164342.7164342.7164342.1980335.877136.593435.67963Total Resource Cost3315331533153345323532613237Processing Time1361113611136111864184919062451Computational Time43.486634.000241.05656.0564383.4802661.8143798.3901Table 3.c : Resource -20 Job -60FCFSSJFSSACOAWPFAWOPFSS-GAUtilization118.5099118.5099118.509993.65225104.7856108.682388.59878Economy41.8652741.8652741.8652738.2427739.75340.18437.25771Total Resource Cost3759375937593887375737723749Processing Time1933019330193302384244125503063Computational Time48.957760.680244.14748.5887415.6853696.6445850.4797Table 3.d : Resource -20 Job -70 .Maximizing Computational Profit in Grid Resource Allocation using Dynamic AlgorithmFCFSSJFSSACOAWPFAWOPFSS-GAUtilization103.6847103.684787.55402102.5304103.6585101.451997.70473Economy39.6838139.6838136.8093339.5081439.281138.551938.061Total Resource Cost3777377739573788373338193726Processing Time1546615466147062007227923452427Computational Time78.717452.71252.218411.8828385.6945651.1487854.7021013Table 4.d : Resource -25 Job -702 YearUtilizationFCFS 91.76289SJF 100.2523SS 100.2523ACO 101.4468AWPF 99.2599AWOPF 100.1643SS-GA 98.619826Economy41.3992542.8809342.8809341.096541.42841.409241.09644( D D D B D ) Volume XIII Issue II Version IUtilization Economy Total Resource Cost Processing Time Total Resource Cost Processing Time Computational TimeFCFS 114.3678 42.94165 2232 2297 4499 22542 37.1361SJF 114.3678 42.94165 2232 2297 4389 23287 40.0114 Table 4.e : Resource -25 Job -80 SS ACO 114.3678 101.8771 42.94165 41.22076 2232 2186 2297 836 4389 4375 23287 2602 50.1337 15.8017AWPF 102.0942 35.8982 2191 842 4356 2672 444.0259AWOPF 104.5534 35.8967 2234 879 4399 2764 787.7263SS-GA 98.84106 40.73676 1139 2184 4343 3185 904.6483Global Journal of Computer Science and TechnologyComputational Time Utilization Economy Total Resource Cost Processing Time Computational Time Utilization52.7427 FCFS 115.3724 41.32471 2731 5594 65.6896 FCFS 98.7189945.809 Table 4.a : Resource -25 Job -40 75.3122 8.6043 SJF SS ACO 115.3724 87.60529 102.2472 41.32471 36.86441 39.51829 2731 2831 2712 5594 5212 1111 47.8802 67.6473 7.7583 Table 4.b : Resource -25 Job -50 SJF SS ACO 99.54338 98.71899 102.9516284.6685 AWPF 106.1026 43.0998 2680 1259 313.8459 AWPF 102.1378568.168 AWOPF 108.8235 43.1747 2710 1257 593.0732 AWOPF 97.3184710.2946 SS-GA 95.41284 38.38091 2663 1666 763.5657 SS-GA 97.31663Economy38.5486339.1982238.5486339.7364537.829136.84837.1611Total Resource Cost3311327633113247328233783238Processing Time100169418100161555161316271881Computational Time51.926956.584149.94611.9545319.5781585.6896775.5075Table 4.c : Resource -25 Job -60© 2013 Global Journals Inc. (US) .FCFSSJFSSACOAWPFAWOPFSS-GAUtilization90.2491890.2491882.23188.0549796.97399.236685.87629Economy34.4963534.4963532.7222634.0429738.013438.196533.56988Total Resource Cost3423342334133346324532833340Processing Time6801680170591277145915571713Computational Time56.698460.83953.809712.8454353.1478649.9953807.4937Table 5.c : Resource -30 Job -60UtilizationFCFS 123.7695SJF 123.7695SS 123.7695ACO 102.6892AWPF 96.4486AWOPF 96.8692SS-GA 97.63258YearEconomy42.9938842.9938842.9938840.3106937.264437.785839.49474Total Resource Cost3733373337333804376138103756Processing Time1256612566130791799196619342348Computational Time52.738745.291858.181220.3609378.1919647.6226860.2208Table 5.d : Resource -30 Job -70FCFSSJFSSACOAWPFAWOPFSS-GAUtilization118.8993113.0306113.030691.6733497.126495.833390.08655Economy40.8160440.105240.105236.7500739.843839.492236.43731Total Resource Cost4363431343134349431243854301( D D D D D D D D )Processing Time1792518260182602199228223833013Computational TimeFCFS 76.1423SJF 84.2506SS 68.7347ACO 19.757AWPF 479.4874AWOPF 843.1664SS-GA 918.3843Utilization88.854103.4735103.473594.64883105.517299.679590.56Economy35.8106238.3931138.3931136.929741.451339.880836.15495Total Resource Cost2237214721472198213022192125Processing Time118013041304724801880914Computational Time60.862551.83657.449814.2908216.1684481.3565729.9444Table 5.a : Resource -30 Job -40FCFSSJFSSACOAWPFAWOPFSS-GAUtilization105.7991102.8886105.799197.26918102.6243101.756896.76585Economy42.2075241.7805342.2075240.8838640.30539.793340.79846Total Resource Cost2727273727272769274027962723Processing Time4040358640401057108111001372Computational Time105.439982.642456.946313.9463286.4863602.7403776.9802Table 5.b : Resource -30 Job -50 5e : Resource -30 Job -80 © 2013 Global Journals Inc. (US) © 2013 Global Journals Inc. (US) Global Journal of Computer Science and Technology * Grid Computing Technology GeorgianaMarin Database Systems Journal 2 3 2011 * Grid Computing With Globus: An Overview and Research Challenges RajeevWankar International Journal of Computer Science and Applications 5 3 2008 * Genetic Algorithm Based Schedulers For Grid Computing Systems AjithAbraham FatosXhafa JavierCarretero International Journal of Innovative Computing, Information and Control 3 5 2007 * Modified Ant Colony Algorithm for Grid Scheduling Mathiyalagan SivanandamSuriya International Journal on Computer Science and Engineering 2 2 2010 * A New Framework for Job Scheduling and Resource Management in Grid Environment AkashMalhotra KunalGupta International Journal of Emerging Technology and Advanced Engineering 2 2012 * A Taxonomy of Grid Resource Selection Mechanisms AdilYousif AbdulHananAbdullah MuhammadShafie International Journal of Grid and Distributed Computing 4 3 2011 Abd Latiff, Mohammed Bakri Bashir * Grid Resource Allocation: A Review NordinBHaruna Ahmed Abba NazleeniZakaria Haron Research Journal of Information Technology 4 2012 * Intelligent Methods for Resource Allocation in Grid Computing HarshBansal BabitaPandey KewalKrishan International Journal of Computer Applications 47 6 2012 * Process Resource Allocation in Grid Computing using Priority Scheduler MayankKumar AbhayMaheshwari Bansal International Journal of Computer Applications 46 11 2012 * A Dynamic Load Balancing Algorithm in Computational Grid Using Fair Scheduling KarthickKumar International Journal of Computer Science Issues 8 1 2011 * An Economic Allocation of Resources for Divisible Workloads in Grid Computing Paradigm ChellappanMurugesan European Journal of Scientific Research 65 3 2011 * Association Based Grid Resource Allocation Algorithm DuraiswamyManavalasundaram European Journal of Scientific Research 78 2 2012