Study Of Cloud Computing Resource Management Algorithm Based On Improved Genetic Algorithm

Cloud computing is a new type of on-demand resource pool can be configured to provide users with computing, networking, storage, and other virtual resources. As a technology providing users with commercial service, rational scheduling system resources is the key to cloud computing. Because of its heterogeneous, dynamic, mass and other characteristics, developers should consider how to reasonably schedule the resources and allow users to access resources in a short time. At the same time, improving resource utilization and reducing energy consumption as much as possible are also urgent problems need to be solved.In this paper, a resource scheduling algorithm based on genetic algorithm is proposed. Genetic Algorithm(GA) is a global optimization search algorithm with a randomized characteristics, it appears based on the survival of the fittest laws of the evolution. Because of its overall search strategy and optimization search method does not depend on other auxiliary knowledge while working, the genetic algorithm has a strong versatility. In addition, it has an excellent performance in solving NP problems, hence many studies have implemented genetic algorithm in solving the problem in resource scheduling of large-scale cluster. In order to save energy and to maximize the economic benefits on the basis of satisfying users' needs, this thesis introduces economic benefits constraints, SLA(Service Level Agreement) constraints and energy constrains in the fitness function, so that the most suitable placement strategy could be found when creating the virtual machine in the physical machine.In order to solve the problem of premature in GA caused by its search strategy, Tabu Search Algorithm(TS) is applied in this thesis to address this issue. TS is a gradual global optimization algorithm that simulate human intelligence developing process. In this algorithm, a good initial solution can greatly improve its searching efficiency. Luckily, the solution obtained by GA can provide such initial solution. As a result, the combination of TS with genetic algorithm can dramatically improve the performance. In this thesis, we will determined if the GA enters premature stage during the calculation process and introduce TS algorithm using genetic algorithm solution as its initial value. While jumping out of local optimal, TS will also produce a new neighborhood to ensure the understanding of diversity, so that the gradually optimized results will be obtained.The cloud simulation tool Clous Sim is introduced in this thesis, based on which the proposed algorithm is simulated. Round-Robin algorithm(RR) and Random Allocation algorithm(RA) are also tested in this work to make a fair comparison with proposed algorithm. The results show that the improved genetic algorithm can effectively allocate the resources in cloud computing and tradeoff within multiple constraints such as SLA constraint and energy constraint to obtain a higher economic efficiency.