Uzma Nisar

Task Scheduling using Hybrid improved crow particle swam optimization algorithm (HICPSOA) Uzma Nisar , Dr. Munam Ali Shah Department of Computer Science COMSATS University Islamabad Islamabad, Pakistan-,-Abstract — Cloud computing allows services to be supplied quickly and managed efficiently over the Internet. Cloud computing allows users to access data and storage services regardless of where they are physically located. Large number of jobs and computer resources in the environment of cloud are available. The task scheduling algorithms, which are crucial to the cloud computing process, find the best virtual machine (VM) assigning resources to complete any given job. Approaches to task scheduling increase the timeliness of the schedule while also saving money. Aim of this paper is to find optimal solution for task scheduling problem in cloud computing environment by utilizing dynamic task scheduling techniques .We solved this problem by using an extended form of CSO known as Improved Crow Search Optimization (ICSO) and another dynamic technique called Particle Swarm Optimization (PSO). ICSO is affective stochastic optimization technique, which has a few distinguished functions. The individuals of the population can be operated in search area simultaneously, and it utilizes probabilistic transition regulations rather than deterministic ones. It additionally has the potential of getting out local minima. Honestly, the benefit of one set of rules can be the remedy for the lack of PSO. The incorporation of ICSO into PSO as a local development technique permits these of rules to maintain the population diversity and escape from nearby optima Keywords—Cloud Computing; Improved Crow Search; Task Scheduling; Algorithm for resource allocation; Particle Swarm Optimization. I. INTRODUCTION As a metaphor for the Internet, cloud computing can be viewed as removing the physical barriers that hinder access to resources while simultaneously providing access to resources from anywhere with internet connectivity [1]. One of important task in cloud computing is task scheduling. A number of task scheduling algorithms are available with their capabilities. For better utilization of resources there is need to choose best of them. For this purpose researchers used many swarm intelligence algorithm and their hybrid forms. One of them is Crow search algorithm (CSA) which is developed by observing the behavior of crow bird. Another swarm intelligence algorithm is Particle Swarm Optimization algorithm (PSO) which is being used in task scheduling in many field of computer sciences. Cloud computing is a new cloud or Internet-based computing model that allows users to get services on demand. Cloud computing allows service providers to give their services to several users at the same time. Cloud computing also allows users to use it as an Internet-based pay-as-you-go service, providing them with services that are built on uncertainty, dynamism, and flexibility. Many clients in the - industries of education, government services, telecommunications, and financial services use it extensively [2]. Because of the tremendous development in demand for cloud computing services and the heterogeneous nature of cloud computing resources, resource scheduling has become increasingly important. Internet users can now access material from anywhere and without having to worry about the hosting infrastructure at any moment. The hosting infrastructure, which consists of several computers with varying capacity, is maintained and managed by the service provider. Cloud computing expands the potential of infrastructure that has Internet connectivity. Cloud service providers profit on the services they supply to cloud service users [3]. In order to reduce reaction task scheduling policy in such systems focuses on time and processing time how to schedule jobs in a way that reduces data transfer and increases the processing ability of these systems, hence improving performance. Scientific workflow applications typically require not only high-performance computation but also large amounts of storage. Grid systems are now used to install popular scientific procedures because to their fast performance and enormous storage capacity. Grid computing, on the other hand, is best suited for specialized applications and not available to consumers for whole world. Cloud computing is a new distributed computing paradigm [4]. Cloud computing is known as latest technology that emerged The term "grid computing" comes from the terms "distributed computing" and "grid computing" it refers to the use of computer resources (hardware, software, and platforms) as a service and making them accessible to clients on demand through the Internet. It is the first technology to make commercial computer science accessible to the general people. It is built on a virtualization strategy that allows users to share resources [5]. Cloud computing can give high performance by evenly and effectively distributing workloads across all resources, resulting in minimal waiting time, optimal resource utilization, maximum throughput, and Execution time. Even still, there remain significant roadblocks. Load balance and Task Scheduling are the most significant because they are the primary determinants of other performance parameters including availability, scalability, and power consumption. As a result, cloud users can provide their consumers with more dependable, accessible, and latest services [6]. The cloud is made up of actual equipment that are housed in cloud service providers' data centers On top of these actual systems, virtualization is supplied. These virtual computers are accessible to cloud users. Different cloud service providers provide various levels of abstraction in their cloud services. Customers may handle incredibly low-level information on Amazon EC2, for example, whereas Google App-Engine is a platform on which programmers may build their applications. Cloud service categorize in 3 platforms: platform as a service, software as a service and infrastructure as a service. These services are access by the Internet from anywhere in globe, with cloud serving for all clients as a single point of contact [7]. The job scheduling approach used in cloud computing environment which has unique impact on its performance. Scheduling in the cloud is the act of allocating VM with resources available to satisfy user requirements or mapping a set received workloads to group a number of virtual computers capable of carrying out the tasks. Resource optimization might be achieved by effective scheduling, which would increase performance and efficiency of the cloud computing environment. Because users are increasingly turning to the cloud for their application needs, effective cloud scheduling algorithms are needed to efficiently allocate user tasks or jobs to appropriate cloud data centers. 1.1 Crow Search Algorithm (CSA): The intelligence of crows (crow species) is well known. When the brain-to-body ratio is taken into account, the crow's brain trails behind the human brain by a little margin. The intellectual traits of crows have been well proven, with several proofs. Despite the fact that they are less intelligent than humans, they manufacture tools and have the capacity to recognize themselves in mirrors. They can recognize each other's faces and warn each other when a threat approaches. They have the most advanced form of communication and can remember where they buried their food even after months have passed. Crows keep a watch on other bird species, looking for places where they could be hiding their food, and then stealing it when the opportunity arises. When a crow commits a theft, it begins to flee to newer sites in order to avoid becoming a victim of theft in the future. They were able to discern the intentions of other criminals since they had committed a robbery themselves, and they were able to protect themselves from being stolen. This behavior has been updated to include crows' population-based behavior, such as storing excess food in concealed spots and recovering it when they are in need [21]. The following are the principles of the Crow Search algorithm: 1. 2. 3. 4. Crows congregate in groups. Crows have the ability to remember where they discarded their food. Crows follow each other and take each other's food whenever they have the chance. Crows defend their territory from intruders with a probability between [0,1] Figure 1: Crow search algorithm II. RELATED WORK: In [8] a work scheduling algorithm had been proposed. In their work, the author’s evaluated four criteria that are inherently conflicting, task transfer time, i.e. task execution cost, task queue length, and power consumption. This algorithm depends on the costs and power usage when execution is done, reducing both from the customer and provider's perspectives. Other multiobjective algorithms such as Multi-Objective Particle Swarm Optimization and MultiObjective Genetic Algorithm were used to find the best solution compare the performance of the proposed algorithm. The acquired results demonstrated the superiority of the proposed approach. A hybrid algorithm that joins Cuckoo Search and Harmony based Search are cloud-based techniques for enhancing scheduling performance in [9] developed CHSA. To carry out intelligent process scheduling, the Cuckoo and Harmony search algorithms were combined.The authors presented a multi-objective function with factors such as energy consumption, cost, credit obtained, memory expended, and penalty accrued. The suggested CHSA's performance was compared to that of hybrid search algorithm, Harmony search and Cuckoo search algorithms in terms of categorize multiobjective parameters. To solve resource scheduling challenges we have cloud computing, a unique multi-objective Cuckoo Search Optimization technique was presented. By reducing the make span, the proposed technique reduced the cost borne by the cloud customer while also improving the system's performance in [10]. This method ensured maximum resource use, resulting in higher profits for cloud service providers. In [11] presented the CPSO algorithm, which combines the Cuckoo Search and PSOA to improve scheduling performance in cloud computing environments. The suggested CPSO algorithm substantially lowered QoS metrics such as makespan, cost, and deadline violation rate. The Cloudsim toolkit was used to test the suggested CPSO algorithm's performance. The proposed CPSO algorithm's efficiency was demonstrated by simulation results. The work was assigned to a virtual machine in [12] that had the lowest execution cost and could fulfil the deadline limitations. In addition, the task-assigned virtual machine was assigned to the most frequently used physical host category, which corresponded to its capabilities. By comparing it to previous multi optimization techniques, the authors assessed the suggested co-optimization methodology's performance, collaborative task scheduling, and VM placement (JTSVMP). The authors of [13] proposed a Hybrid Electro Search with a Genetic Algorithm (HESGA) for merging the genetic algorithm with the electro search algorithm to optimize the cloud-based work scheduling procedure To fine-tune the task scheduling activity, the authors employed QoS criteria such as makespan, load balancing, resource utilization, and cost. The Genetic Method yielded the local and best optimum answers, whereas the Electro Search method yielded the best global optimal results. We suggested HESGA outperformed a variety of other current algorithms, according to experimental data. By integrating their most attractive properties, [14] has merged the Bacterial Foraging (BF) algorithms and the genetic algorithm (GA). As a result, scheduling performance in the cloud environment has improved. With regard to both economic and ecological viewpoints, the proposed hybridized scheduling method was successful in compensating for the reduction in makespan and energy usage. When the preceding literature analysis is combined, it does not give a nearest best fine solution for the QoS which depends on parameters and give the best cost and makespan. The suggested Crow Algorithm (HICPSOA ) takes into In order to optimize resource utilization and work scheduling activity among virtual machines in the cloud, take into consideration the makespan and cost parameters. [15, 16]. Reference [8] Abstract Multi-objectiveoriented scheduling for clouds [9] A Hybrid Approach for Task Scheduling in Cloud Computing Environment [10] A multiobjective optimal task scheduling in cloud environment Hybrid electro search scheduling in cloud computing [13] Techniques MultiObjective PSO and MultiObjective GA Cuckoo search and harmony based search Contribution Scheduling algorithm proposed CSO Technique Reduce Make span and improve system performance Optimized the cloud base work Hybrid Electro Search with a Genetic Algorithm (HESGA) Table No: 1 (Comparison) Enhancing scheduling performance III. PROBLEM WITH EXISTING FRAMEWORK: The cloud service providers provide a cloud environment consisting of a physical machine (PM) and a virtual machine (VM) for their customers allowing for a public interface Customers who use the cloud using the UI, they may submit their jobs. All of these are examples of the job requests that have been received are compiled and used properly by the Request Manager. The Resource Monitor maintains track of the cloud resource pool's availability, which includes CPU, memory, and storage. The Scheduler component successfully organizes work in the cloud environment in order to minimize the specified fitness function. The virtual machines are allocated limited tasks based on their performance throughout the scheduling process. After acquiring the necessary information from both the Request Manager and the Resource Monitor, the Scheduler begins scheduling the jobs. After gathering the necessary data, a judgement is made on job assignment to appropriate virtual machines. Each job must be assigned to the proper virtual machine. After acquiring the VMs' location information, the allocation procedure may be fine-tuned. This information aids in the reduction of a variety of aspects, including relocation costs. Existing solution used crow search algorithm (CSA) which have issue of premature convergence. As a result of the unproductive exploration of the search technique, convergence is not guaranteed [19]. IV. PROPOSED METHODOLOGY: In general, the proposed approach is aimed at optimizing task scheduling by incorporating Particle Swarm Optimization (PSO) and Improved Crow Search Optimization (ICSO) algorithms. Crow Search Algorithm (CSA) was developed to solve several optimization problems using swarm intelligence meta-heuristics. The basic concept behind this algorithm draws inspiration from the memory ability, communication skill, and social behavior of crows [20]. In a recent survey on Crow Search Mr. Meraihi et al. described different variants of CSA including its modified versions and hybridized versions. One of modified version is Dynamic Crow Search Algorithm [21]. Mr. Sahoo et al. [22] proposed Improved Crow Search Algorithm (ICSA) for task of scheduling in multiprocessing environment and Mr. Cuveas et al. [23] proposed this for the task of optimal capacitor allocation. ICSA outperformed on basic Crow Search Algorithm [22]. A PSO model was developed based on these five principles by Kennedy and Eberhart. PSO uses a swarm intelligence approach to arrive at a solution. PSO is, therefore, a swarm-based intelligent search algorithm. Random solutions are used to perform this search. Each individual potential solution is known as a particle, and this collection of potential solutions is known as a swarm [21]. A short introduction to Task Assignment Problems based on ICSO is provided in this section. The schedule is initially selected from a population of N flocks and a d-dimensional environment in order to develop the ICSO approaches. Assigning the tasks to the processors is based on the number of flocks in each schedule. Problems of task assignment are formulated as optimization problems of functions whose purpose is to achieve the lowest possible cost. Figure 3 shows the algorithm for assigning tasks by ICSO. Figure 2: Proposed Methodology 4.1 Improved Crow Search Optimization (ICSO): Existing solution used crow search algorithm (CSO) which have issue of premature convergence. As a result of the unproductive exploration of the search technique, convergence is not guaranteed [19]. According to the classical CSO, most of the search process is determined by random movements (evasion), and awareness probabilities (AP). According to the ICSO method, the classical CSO method is used, but the random movement and the AP are recalculated as Dynamic AP (DAP) and Levy flight.The equation for the Dynamic AP is as follows (Eq. 1) from [24]: Eq. 1 Figure 3: Flowchart of ICSO A value of 'wv' signifies a bad objective value. Taking this into account, we calculate levy flights as follows (Eq. 2) [24]: Eq. 2 Zi represents the levy distribution and Cbest represents the best solution so far. As a result, Cit+1 is calculated as follows (Eq. 3)[24]: Eq. 3 4.1.1 Task assignment using ICSO: 4.2 Particle Swarm Optimization (PSO): A particle in PSO learns in two different ways. During movement, particles learn both from other particles and from their own experiences. While learning from others is referred to as social learning, learning from experiences is referred to as cognitive learning. The particle gbest stores the best solution that was visited by each particle in its memory as a result of social learning. Cognitive learning leads to the particle storing in memory the best solution it visited so far, called pbest. Whenever a particle changes direction or magnitude, a factor called velocity determines the changes. Velocities are the rates at which positions change with time. Where time is the iteration of the PSO. Therefore, PSO can be defined as the change in position within an iteration. By increasing the iteration counter by unity, the velocity v and position x become equal. Following (Eq. 4) equation is used to update the particle's velocity[25]. Eq. 4 A position update equation is used to update the position (Eq. 5)[25] Eq. 5 This equation has d = 1, 2,..., D representing the dimension and i = 1, 2,..., S representing the particle index. S represents the size of the swarm, and c1 and c2 represent constants called cognitive and social scaling parameters, respectively or simply acceleration coefficients [25]. 4.2.1 Task assignment using PSO: Observe in Eq. 4 and Eq. 5 that the dimensions of each particle in the problem space are updated independently of one another. The only link between the dimensions is introduced by the objective function, i.e., by the locations of the best positions found so far gbest and pbest. PSO algorithm is defined by Eq. 4 and Eq. 5. Figure 5 describes our algorithmic approach to PSO. 4.3 Hybrid ICSO and PSO: We proposed hybrid form of Improved Crow Search Optimization (ICSO) and Particle Swarm Optimization (PSO) algorithm for the purpose of task scheduling in cloud computing environment. The social and cognitive version makes PSO awareness greater at the cooperation some of the debris. inside the original PSO, all debris inside the swarm examine from the gbest even if the modern gbest is some distance from the global most suitable, and within the local paradigm, debris have statistics only of their own and their nearest array neighbors’ bests, rather than that of entire organization. In the meantime, particles inside the swarm willing no way be removed even supposing they are impossible to reach the pleasant role, which can also waste the restricted computational assets. Therefore, after some generations, the population diversity might be substantially reduced, and the PSO might lead to a untimely convergence to a neighborhood most appropriate. ICSO is affective stochastic optimization technique, which has a few distinguished functions. The individuals of the population can be operated in search area simultaneously, and it utilizes probabilistic transition regulations rather than deterministic ones. It additionally has the potential of getting out local minima. Honestly, the benefit of one set of rules can be the remedy for the lack of PSO. The incorporation of ICSo into PSO as a local development technique permits these of rules to maintain the population diversity and escape from nearby optima. 4.3.1 Task assignment using Hybrid ICSO and PSO: The Particle and Crow algorithms were adapted to generate the job permutation, and the permutation was then mapped to the position values using the SPV rule to ensure that the initial population achieved high-quality solutions with high diversity. Furthermore, we only generate the position value for crows whose velocity is unknown. On the other hand, to reduce the search space, we randomize the initial population using Eq. 6 and Eq. b given below[26]: Eq. 6 Eq. 7 In the above equations both 𝑣min and are −2, 𝑣𝑚𝑎𝑥 𝑥𝑚𝑎𝑥 are 2, and 𝑟1 and 𝑟2 are random numbers of interval [0,1]. A hybrid operator is being used at some particular iterations. Some of the individuals are chosen from ICSO and PSO and swapped by using probabilities based fitness valued, a method presented by Goldberg [26]. The technique of roulette given below in Eq. 8. Figure 4: Flowchart of Particle Swam Optimization experiments were conducted by varying the input task from 50 to 300 numbers. The performance of the proposed Eq. 8 At start each particle (crow) is randomly selected. Then from NEH algorithm output a solution is generated. Both ICSO and PSO is simultaneously executed. The hybrid parameter is utilize at some random iterations for example 20 iterations to swap to particle. Implementation detailed of proposed hybrid ICSO and PSO is given in Figure 5 Entity Type Task (Cloudlet) Virtual Machine Parameters Task Length Total Tasks File size MIPS Number of VM Bandwidth Memory Storage Unity cost of VM Values- Table No 2: parameter setting for Cloud Simulator Hybrid (ICPSOA) had been evaluated by comparing its results with respect to cost and execution time (makespan) parameters against MO-ACO, ACO and MIN-MIN algorithms. Both ACO and Min-Min algorithms aim to provide optimized solutions for task scheduling activity in cloud by continuously iterating the candidate solutions. In a similar manner, the MO-ACO algorithm generates optimized solution for task scheduling activity by taking into consideration of makespan and cost parameters. Since the proposed (HICPSOA) also strives to provide a near optimal solution to task scheduling activity, it has been compared with the MO-ACO, ACO, Min-Min algorithms. The Table 2 lists out the types of entity, parameters and their corresponding values that are considered in the experiment. 5.1 Comparison of makespan: Figure 5: Flowchart of HICPSOA V. RESULTS AND CONCLUSIONS: Experiments were conducted in a simulated environment using Java (jdk 1.6) with Cloudsim tool. The implementation setup comprised of a PC with Windows 7 OS @ 2 GHz dual core, a RAM of 6 GB and a 64-bit windows 2008 OS. The The proposed Hybrid (ICPSOA) increases the efficiency of the task scheduling process by minimizing the makespan and cost parameters. The performance of the proposed HICPSOA with respect to the makespan parameter had been compared with MO-ACO, ACO and MIN-MIN algorithms. In Figure 6, the comparison of task execution time using 10 VMs by applying HICPSOA , MO-ACO, ACO and MIN-MIN algorithms had been shown. For 50 number of tasks, the makespan values obtained are 130,165, 174.69 and 183.706 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. When the task count is increased to 100 numbers, the makespan values obtained are 330, 332.347, 358.802 and385.199 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. For 150 tasks, the makespan values obtained are 511.4, 516.45, 542.876 and 560.606 for HICPSOA, MOACO, ACO and MIN-MIN respectively. For 200 tasks, the values obtained are 660.8, 674.495, 726.978 and 770.794 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. For 250 tasks, the values obtained are 795.3, 797.702, 850.157 and 981.011 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. On the whole it could be inferred that the proposed HICPSOA is superior to other algorithms while comparing For Figure 7 using 20 VMs by applying HICPSOA, MOACO, ACO and MIN-MIN algorithms had been shown. For 50 number of tasks, the makespan values obtained are 70.2, 80.8, 86.5 and 90.5 for HICPSOA, MO-ACO, ACO and MINMIN respectively. When the task count is increased to 100 numbers, the makespan values obtained are 151.1, 165.4, 178.5 and 191.5 for HICPSOA, MO-ACO, ACO and MINMIN respectively. For 150 tasks, the values obtained are 237.1, 256.4,265 and 270.7for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. For 200 tasks, the values obtained are 314, 330.5, 356 and 375 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. For 250 tasks, the values obtained are 378.1, 395.7, 423.5 and 488.445 for HICPSOA, MOACO, ACO and MIN-MIN respectively. Hence it could be inferred that the proposed HICPSOA is superior to other algorithms while comparing the makespan values deduced using 20 VMs. 15000.2, 6284.7, 17986.1 and 19930.6 for MO-ACO, ACO and MIN-MIN respectively. For 200 number of tasks, the cost values are 22648.9, 22847.2, 24305.6 and 26250 for MOACO, ACO and MIN-MIN respectively. For 250 number of tasks, the cost values are 27500, 28437.5, 30381.9 and 33541.7 for MO-ACO, ACO and MINMIN respectively. On the whole it could be inferred that the proposed HICPSOA is superior to other algorithms while comparing the cost values deduced using 10 VMs. Figure 8 shows the cost expended for tasks numbered 50, 100, 150, 200 and 250 using 20 VMs by applying the HICPSOA, MO-ACO, ACO and MIN-MIN algorithms. It could be seen that cost values obtained for 50 tasks are 2703.4, 2793.34, 2925.32 and 2840.67 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. The cost values in the case of 100 tasks are 4855.7, 5468.75,5987.4 and 6452.89 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. In the case of 150 tasks, the values obtained are 7486.54, 8142.35, 8993.05 and 9978.2 for MO-ACO, ACO and MIN-MIN respectively. For 200 number of tasks, the cost values are-, 11420.8, 12158.9 and 13128 for MO-ACO, ACO and MINMIN respectively. For 250 number of tasks, the cost values are 13579, 14325.9, 15296 and- for MO-ACO, ACO and MIN-MIN respectively. Hence it could be inferred that the proposed HICPSOA is superior to other algorithms while comparing the cost values deduced using 20 VMs. Figure 6: Makespan of 10VMs Figure 8: Cost of 10VMs Figure 7: Makespan of 20VMs 5.2 Comparison of Cost: The performance of the proposed HICPSOA with respect to the cost parameter had been compared with MO-ACO, ACO and MIN-MIN algorithms. Figure 8 shows the cost expended for tasks numbered 50, 100, 150, 200 and 250 using 10 VMs by applying the HICPSOA, MO-ACO, ACO and MIN-MIN algorithms. It could be seen that cost values obtained for 50 tasks are-, 5590.28, 5833.3 and 5900.1 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. The cost values in the case of 100 tasks are 5400, 5590.28, 5833.3 and 5900.1 for HICPSOA, MO-ACO, ACO and MIN-MIN respectively. In the case of 150 tasks, the values obtained are Figure 9: Cost of 20VMs VI. CONCLUSION: As a metaphor for the Internet, cloud computing can be viewed as removing the physical barriers that hinder access to resources while simultaneously providing access to resources from anywhere with internet connectivity. The task scheduling algorithms, which are crucial to the cloud computing process, find the best virtual machine (VM) assigning resources to complete any given job. Approaches to task scheduling increase the timeliness of the schedule while also saving money. The Crow Swarming Algorithm (CSA) simulates crow behavior concerning food storage and retrieval by using swarm intelligence. Among all food locations, the location where the most food is stored is considered to be the global optimal solution in optimization theory. The crow is the searcher, and the surrounding environment is the search space. CSA simulates crow behavior to find optimal solutions to a variety of optimization problems. It has gained wide interest due to its simplicity, simplicity of implementation, flexibility, and efficiency. An optimization problem that arises in many science and engineering fields is solved by a swarm-intelligent algorithm, inspired from a bird's flock or a fish's schooling, which solves nonlinear, nonconvex or combinatorial optimization problems. This paper aim to optimize task scheduling in cloud computing environment by utilizing dynamic task scheduling techniques. In cloud computing task assignment is an optimization problem. We solved this problem by using an extended form of CSA known as Improved Crow Search Optimization (ICSO) and another dynamic technique called Particle Swarm Optimization (PSO). VII. REFERENCES: 1- Ibrahim, I. M. (2021). Task scheduling algorithms in cloud computing: A review. Turkish Journal of Computer and Mathematics Education (TURCOMAT), 12(4),-.Saleh, I. A., Alsaif, O. I., Muhamed, S. A., & Essa, E. I. (2019, September). Task scheduling for cloud computing based on firefly algorithm. In Journal of Physics: Conference Series (Vol. 1294, No. 4, p. 042004). IOP Publishing. 2- Houssein, E. H., Gad, A. G., Wazery, Y. M., & Suganthan, P. N. (2021). Task scheduling in cloud computing based on meta-heuristics: Review, taxonomy, open challenges, and future trends. Swarm and Evolutionary Computation,-- Beegom, A. A., & Rajasree, M. S. (2019). Integer-pso: a discrete pso algorithm for task scheduling in cloud computing systems. Evolutionary Intelligence, 12(2),-- Velliangiri, S., Karthikeyan, P., Xavier, V. A., & Baswaraj, D. (2021). Hybrid electro search with genetic algorithm for task scheduling in cloud computing. Ain Shams Engineering Journal, 12(1),-- Kadhem, A., Aziz, H., Gasim, T. Y., & Ismaeel, H. (2021). Task Scheduling Algorithms in Cloud Computing (No. 6756). EasyChair. 6- Arunarani, A. R., Manjula, D., & Sugumaran, V. (2019). Task scheduling techniques in cloud computing: A literature survey. Future Generation Computer Systems, 91,-- F. Ramezani, J. Lu, J. Taheri, and F. K. Hussain, “Evolutionary algorithm-based multi-objective task scheduling optimization model in cloud environments”, World Wide Web, Vol. 18, No. 6, pp-, 2015. 8- S. H. H. Madni, M. S. Abd Latiff, and J. Ali, “Multiobjective-oriented cuckoo search optimization-based resource scheduling algorithm for clouds”, Arabian Journal for Science and Engineering, Vol. 44, No. 4, pp-, 2019. 9- K. Pradeep and T. P. Jacob, “A Hybrid Approach for Task Scheduling Using the Cuckoo and Harmony Search in Cloud Computing Environment”, Wireless Personal Communications, Vol. 101, No. 4, pp-, 2018. 10- K. Pradeep and T. P. Jacob, “A multi-objective optimal task scheduling in cloud environment using cuckoo particle swarm optimization”, Wireless Personal Communications, Vol. 109, No. 1, pp. 315-331, 2019. 11- A. Dabiah, H. Tianfield, Y. Zhang, and B. Pranggono, “A metaheuristic method for joint task scheduling and virtual machine placement in cloud data centers”, Future Generation Computer Systems, Vol. 115, pp. 201-212, 2021. 12- S. Srichandan, T. A. Kumar, and S. Bibhudatta, “Task scheduling for cloud computing using multi-objective hybrid bacteria foraging algorithm”, Future Computing and Informatics Journal, Vol. 3, No. 2, pp. 210-230, 2018. 13- S. Velliangiri, P. Karthikeyan, V. A. Xavier, and D. Baswaraj, “Hybrid electro search with genetic algorithm for task scheduling in cloud computing”, Ain Shams Engineering Journal, Vol. 12, No. 1, pp. 631-639, 2021. 14- P. Krishnadoss and P. Jacob, “OCSA: Task Scheduling Algorithm in Cloud Computing Environment”, International Journal of Intelligent Engineering and Systems, Vol. 11, No. 3, pp. 271-279, 2018. 15- K. Pradeep and T. P. Jacob, “CGSA scheduler: A multiobjective-based hybrid approach for task scheduling in cloud environment”, Information Security Journal: A Global Perspective, Vol. 27, No. 2, pp. 77-91, 2018. 16- Krishnadoss, P., Pradeep, N., Ali, J., Nanjappan, M., Krishnamoorthy, P., & Kedalu Poornachary, V. (2021). HICPSOA : Hybrid cuckoo crow search algorithm for task scheduling in cloud computing. International Journal of Intelligent Engineering and Systems, 14(4),-- A. Askarzadeh, “A novel metaheuristic method for solving constrained engineering optimization problems”, Crow Search Algorithm Computers & Structures, Vol. 169, pp. 1-12, 2017. 18- N. Gobalakrishnan and C. Arun, “A New MultiObjective Optimal Programming Model for Task Scheduling using Genetic Gray Wolf Optimization in Cloud Computing”, The Computer Journal, Vol. 61, No. 10, pp-, 2018. 19- Dos Santos Coelho, L., Richter, C., Mariani, V. C., & Askarzadeh, A. (2016, November). Modified crow search approach applied to electromagnetic optimization. In 2016 IEEE Conference on Electromagnetic Field Computation (CEFC) (pp. 1-1). IEEE. 20- Askarzadeh, A. (2016). A novel metaheuristic method for solving constrained engineering optimization problems: crow search algorithm. Computers & Structures, 169, 112. 21- Meraihi, Y., Gabis, A. B., Ramdane-Cherif, A., & Acheli, D. (2021). A comprehensive survey of Crow Search Algorithm and its applications. Artificial Intelligence Review, 54(4),-- Sahoo, R. M., & Padhy, S. K. (2019, August). Improved crow search optimization for multiprocessor task scheduling: A novel approach. In International Conference on Application of Robotics in Industry using Advanced Mechanisms (pp. 1-13). Springer, Cham. 23- Cuevas, E., Gálvez, J., & Avalos, O. (2020). An enhanced crow search algorithm applied to energy approaches. In Recent Metaheuristics Algorithms for Parameter Identification (pp. 27-49). Springer, Cham. 24- Díaz, P., Pérez-Cisneros, M., Cuevas, E., Avalos, O., Gálvez, J., Hinojosa, S., & Zaldivar, D. (2018). An improved crow search algorithm applied to energy problems. Energies, 11(3), 571. 25- Kennedy, J., & Eberhart,Particle swarm optimization. In Proceedings of ICNN'95-international conference on neural networks (Vol. 4, pp. 2017). IEEE. 26- Goldberg, D. E., & Holland, J. H. (2018). Genetic algorithms and machine learning. Addison-Wesley Longman Publishing Co., Inc., 2018.