基于遗传算法的负载平衡云计算_java_代码_下载

# Genetic-based-Load-Balancing-Cloud-Computing Abstract— Job scheduling is a very significant aspect of cloud computing. Different kinds of services are provided to the consumers within desirable throughput. A lot of optimization techniques emerged to enhance efficiency and reduce expenses. This paper provides an approach to an online mode heuristic algorithm named as ‘Genetic Algorithm’ and its comparison with the batch mode heuristic algorithms. Cloudsim is used to perform the simulation. The results of the simulation show that the Genetic algorithm performs well compared to First Come First Serve Algorithm, Round Robin Algorithm, and Shortest Job First algorithm. Introduction Cloud Computing is the on-demand delivery of computing resources such as servers, storage, databases, software, networking, analytics, and intelligence over the internet with a pay-per-use model. The users are charged based on their usage hence no additional payment for unused infrastructure. Various virtualization techniques are used for real-time information exchange [4]. There are a few most common Cloud deployments – Public Cloud - In Public cloud, the services are owned and operated by third-party service providers. These services are shared by multiple organizations. Private Cloud – Private cloud refers to resources used by a single organization or business. Services such as servers, networks, and data centers are completely dedicated to one organization. Hybrid Cloud – This type of cloud combines the public and private cloud and allows the data to be shared among them. There are two approaches an organization can follow here – Use the private cloud for some services and public cloud for others. Use the public cloud as a backup of the private cloud. It provides three types of service models – Infrastructure as a Services (IaaS) - It provides services that include servers, networking, storage. It gives the highest extensibility and control over the management of computing resources. The flexible, innovative services are available on demand. It is similar to existing IT resources. Platform as a Services (PaaS) - It is a service to the user that gives a platform to develop or test the application on the cloud. It removes the need for organizations to manage the underlying infrastructure. It reduces the complexity of middleware as a service. Software as a Services (SaaS) - It is a method of delivering software applications over the internet on-demand and on a subscription basis. In this model, the user has not to think about how the service is maintained or how the underlying infrastructure is managed. The user should think only about how to use particular software. Everything else is managed and maintained by the service provider. The demands of the cloud are increasing day-by-day. It has become difficult but necessary to provide the capability of handling the increasing requests. To maintain the efficacy of the performance, there should not be any difference in providing the services even after increasing the number of requirements [2]. Job scheduling is the principal activity of cloud computing. Job scheduling with increasing needs has become the concern of cloud computing since it is very important to provide effective service. It mainly aims to distribute the system load and improve system performance. Achieving a high computing execution, managing efficiency amid jobs, the most desirable system throughput, and reduced expense is the intent of scheduling jobs. To achieve the goals stated above, it needs to allocate a specific job to a particular time and resource. There are two types of heuristic job scheduling algorithms in cloud computing which help provide optimization. Heuristic Algorithms find a solution among all possible options, although it does not guarantee that it is an optimal one – Batch Mode Heuristic Scheduling Algorithm - The jobs are collected in a set and queued. The process of mapping of events begins in which jobs are mapped to a particular virtual machine at predefined time intervals. Online Mode Heuristic Scheduling Algorithm - In this mode, whenever a job comes, the job scheduler assigns it to an available machine. On a cloud, the speed of each processor varies quickly as per the load available at that instance. It is more feasible to apply online mode scheduling algorithm. We are using an online mode heuristic scheduling algorithm called as ‘Genetic Algorithm’. A genetic algorithm is a search heuristic algorithm inspired by Charles Darwin’s theory of natural evolution [5]. It reflects a process of natural selection where the fittest individuals are selected to produce offspring of the next generation. The process starts with the selection of fittest individuals from a population. They produce offspring which inherit the characteristics of parents. This offspring is added to the next generation. If parents have better fitness, then their offspring have a better chance of survival [10]. This process iterates until a generation with the fittest individuals is discovered [9]. We consider a set of solutions for a problem and select the set of best ones. Initialization The Genetic Algorithms require a population of n individuals to generate an optimal solution. Each individual is represented by a chromosome. From the n individual population, some individuals are selected based on some functions and operations are performed on them to find the fittest individual [3]. Fitness Function This determines the productivity of individuals in the population. It represents the superiority and performance of one individual over the other. Low fitness value individuals are not selected for further operations. Fitness function is an important part of the Genetic Algorithm [3]. Selection This is the process where the individuals are selected from the n individual’s population based on a strategy which can be tournament selection, roulette wheel, selection based on rank. This is an intermediate solution for the next generation. Crossover In this operation, two chromosomes are selected and divided into parts then they are combined in such a way that the first part of the first chromosome is combined with the latter part of the second chromosome and vice versa. This improves the searching mechanism in the Genetic algorithm [4]. Mutation This operation is used after the crossover step. If the crossover is not able to generate the required diversity due to crossover between the same populations repeatedly and hence mutation operation is done to bring some variation in the population. Usually, the value of mutation is kept very low. This operation alters the gene value in the chromosome. This helps the Genetic Algorithm to produce even more optimal results [4]. Simulation Simulation is a demand for cloud computing to behold the implementation of real-time outlines [7]. There are various tools available for it such as Cloudsim, CloudAnalyst, iCanCloud, GroudSim, NetworkSim, SPECI, and Cloud Report. We are using Cloudsim and CloudAnalyst in our project. CloudSim Cloudsim is a software which is used to perform modeling, simulation, and experiments in cloud computing services and infrastructures. "The tool is developed by Cloud Computing and Distributed Systems Laboratory at the University of Melbourne." We are using it to implement different job scheduling algorithms such as First Come First Serve algorithm, Round Robin algorithm, Shortest Job First algorithm, and Mainly Genetic algorithm. It can instantiate many data centers which consist of storage servers and physical host machines. These machines host multiple VMs executing several cloudlets. CloudSim can perform simulations of assigning and executing a workload on a cloud infrastructure [7]. Problem Definition Suppose that there are m tasks T={t_1,t_2,,...,t_m