Enhanced Task Scheduling Using Optimized Particle Swarm Optimization Algorithm in Cloud Computing Environment

The most significant constraint in cloud computing infrastructure is the job/task scheduling which affords the vital role of efficiency of the entire cloud computing services and offerings. Job/ task scheduling in cloud infrastructure means that to assign best appropriate cloud resources for the given job/task by considering of different factors: execution time and cost, infrastructure scalability and reliability, platform availability and throughput, resource utilization and makespan. The proposed enhanced task scheduling algorithm using particle swarm optimization considers optimization of makespan and scheduling time. We propose the proposed model by using dynamic adjustment of parameters with discrete positioning (DAPDP) based algorithm to schedule and allocate cloud jobs/tasks that ensues optimized makespan and scheduling time. DAPDP can witness a substantial role in attaining reliability in by seeing the available, scheduled and allocated cloud resources. Our approach DAPDP compared with other existing particle swarm and optimization job/task scheduling algorithms to prove that DAPDP can save in makespan, scheduling and execution time.


Introduction
The Cloud computing paradigm allows on-demand delivery of computing resources-computing capabilities, database management, storage abilities, application support, etc. over a dynamic platform through the Internet connectivity by using the pay-as-you-go model [1][2].With this technology, organizations are released from high investments in software/ hardware procurement, organization, and management [3][4].
In its place, the right kind and size of the computing resources are provisioned from the cloud service provider [5].Cloud computing delivers a simple approach to access the various services-servers, networks, storage, processing, databases, and application services from the vendors over the Internet as shown in Figure 1.

Related work
The NP-hard problem of load balancing is a disadvantageous event on the cloud service provider's side that degrades the cloud performance and efficiency [6].To maintain guaranteed Quality of Service on promised service level agreements between cloud consumers and cloud providers continuous audits takes place.Here, there arises the requirement of an efficient load balancing algorithm which is a robust topic of research among several researchers as given in Table 1.

Proposed model
The load balancing in cloud computing can be done at the physical machine level or VM level.The proposed framework recommends an efficient hybrid meta-heuristic load balancing algorithm to ensure a better quality of service through the optimal use of cloud resources.It is implemented in two main modules.The first module is having two phases.In the first phase, it optimizes the make span and scheduling time of the tasks by using improved particle swarm optimization (PSO) algorithm.In the second phase, it reduces execution costs by optimizing the fitness function of PSO.
In the second module, it achieves improved load prediction for self-adaptive clouds by using a non-linear support vector machine algorithm for better load balancing in a cloud computing environment.
Through the improved fitness and kernel mathematical functions of the proposed model, the data center provides several virtual machines to allocate and execute tasks among them.
Our new load balancing model lets cloud services providers optimize and maximize their resource usage.Algorithm of the proposed dynamic adjustment of parameters with discrete positioning (DAPDP) based particle swarm optimization algorithm is given below.

Step1-Start
Step 2-Initialize DAPDP parameters {Maximum number of iterations as Imax, size of the population as P, inertia weight (ꞷ) range [max-min], learning factors l1 and l2} Step 3-Initialize P, initial positions and particle velocities randomly Step 4-Get fitness value of each particle by using following.
Step 8-Check whether the current iteration reaches the best value, if it is not executing Step-3 for the next iteration and obtain optimal value.
Step 9-Obtain and output optimal results.

Execution Environment and Parameter Setting
Cloud computing offers computing, storage, and software as a service.Through cloud computing provides different types of services over the internet, from software and analytics to secure and safe data storage and networking resources, everything can be delivered through the cloud framework.
To implement the proposed model, the CloudSim-cloud simulator is used to simulate the physical and logical structure of data centers in the cloud which contains thousands of physical machines and hundreds of virtual machines in an ordinary computer.Simulators may vary in functionalities such as availability of graphical interface, basic programming language, extensibility, etc.
CloudSim is a java based multi-layered cloud simulator.CloudSim can be used as a standalone simulator to study and analyze various existing policies and cloud entities also, it is possible to change the policies defined for various cloud operations depending on which class to extend.It provides essential classes for describing data centers, computational resources, virtual machines, applications, users, and policies for managing various parts of the system such as scheduling and provisioning.It can also be used to access the capabilities from various perspectives such as cost, application, execution, time, etc.
It also supports the evaluation of Green IT policies.CloudSim is flexible enough to be used as a library that allows you to add the desired scenario by writing a java program.Parameters of the proposed particle swarm optimization based dynamic adjustment of parameters with discrete positioning (DAPDP) is represented in Table 2.

Results and Discussions
Results of the proposed dynamic adjustment of parameters with discrete positioning (DAPDP) algorithm is compared with the existing state of the art algorithms using makespan and execution time as QoS parameters and are represented in Table 3. Makespan results are compared with Min-Min, Max-Min, QoS guided and PSO algorithms and plotted in Figure 2.

Conclusion and Future Work
In our research, to handle task scheduling problem in cloud computing environment, we proposed dynamic adjustment of parameters with discrete positioning (DAPDP) to satisfy QoS requirements.We improved particle swarm optimization by optimizing the inertia weight and allowing the position of the particle as coding discrete.And also we recommended a task scheduling algorithm to effectively handle task scheduling problem by improving the fairness function to optimize make span, scheduling and execution time.In future, the proposed model can be improved by reducing the execution cost by optimizing the fitness function of PSO with adaptive pbest updating through probability and similarity balanced discretization method.This approach will mark our method more expedient and effective.

Figure 1 .
Figure 1.A primer on cloud computing environment

Table 1 .
State of the art load balancing algorithms in the cloud computing environment

Table 3
Figure 2. Makespan results of proposed algorithms vs existing algorithms