Research On Intelligent Optimization Algorithm In Cloud Computing Environment And Its Application In SaaS

Cloud computing is the mixed evolution and jump result of those concepts such as Virtualization、Utility Computing、IaaS (Infrastructure as a Service)、 PaaS (Platform as a Service)、SaaS (Software as a Service) and so on.SaaS takes service as the essence of software. The user’s demand for software is actually a need for the application service, and the user uses the software is actually the consumption of application service. Compared with the traditional software, SaaS service depends on software and Internet. No matter from the technology or business, it has different characteristics with the traditional software: the Internet characteristics, service characteristics, use on demand, rapid deployment, etc.Now, the research and application about harmony search algorithm and ant colony optimization have made some valuable research results in SaaS, but most of them is on the mode of centralized serial, the study of parallelizing harmony search algorithm and ant colony optimization is few. application service quantity and types, service tenants number and amount of information data in SaaS is growth in geometry curve. One of SaaS is single-function and dynamically changes; it cannot meet the increasingly complex needs of users. So the demand to the service in SaaS platform is more and more intelligent.For this purpose, this dissertation has a pertinence and depth study to intelligent optimization algorithm in cloud computing environment and its application in SaaS. The dissertation mainly studies the contents:improved harmony search algorithm and harmony search algorithm based on MapReduce, cloud harmony search algorithms for the composite SaaS placement problem, ant colony optimization based on MapReduce, Service dynamic selection algorithm based on MapReduce and ant colony optimization in SaaS and so on. The main research work and innovative points are as follows:1. improved harmony search algorithm and harmony search algorithm based on MapReduceIn this dissertation, the basic harmony search algorithm is improved and the improved harmony search algorithm、Multi-objective improved harmony search algorithm、the harmony search algorithm based on MapReduce are put forward.Improved the basic cloud harmony search algorithm, using the method of Skyline to initialize harmony memory, it could improve the algorithm efficiency. And using the method of TOPSIS to choose the knowledge as a service (KaaS), it ensures the effectiveness of the solution.In this dissertation, the KaaS combinatorial optimization problem in Hadoop is solved. An cloud harmony search algorithm is proposed by using HDFS and MapReduce to paralleling harmony search algorithm, defining the operation of Map and Reduce.2. cloud harmony search algorithm for the composite SaaS placement problemIn this dissertation, the problems in the SaaS practical application are analyzed, the mathematical model of SaaS placement problem is established, and the cloud harmony search algorithm is proposed for solving the problem. In hadoop, we define the operation of Map and Reduce about cloud harmony search algorithm, at the same time we improve it so that the algorithm efficiency is increased and the calculation speed is accelerated. The result of simulation experiment illustrate that improved cloud harmony search algorithm can solve SaaS placement problem effectively and the results are relatively ideal.3. Ant colony optimization based on MapReduceIn this dissertation, ant colony optimization based on MapReduce is presented. This algorithm can give full play to characteristics of distributed and parallel in cloud computing. Making full use of cloud computing powerful computing and storage capacity, it provides a novel idea and a new method, and promotes the intelligent development of cloud computing.An improved ant colony optimization based on MapReduce is proposed in this dissertation,. It mixes the thought of dividing and conquering and the simulated annealing algorithm with the cloud ant colony optimization to compensate for the defects. And it is applied to the solving of TSP for Verifying the feasibility of the algorithm, validity, convergence, scalability and time consuming.4. Service dynamic selection algorithm based on MapReduce and ant colony optimization in SaaSA service dynamic selection algorithm based on MapReduce and Multi-objective ant colony optimization in SaaS is put forward in this dissertation. The algorithm adopts semi-auto mode and the global optimal multi-objective services selection optimization scheme based on QoS to study the large-scale service dynamic selection problem. It parallelizes ant colony optimization using MapReduce and distributed file system, Designs the multi-objective ant colony optimization, and at the same time optimizes multiobject parameters, Eventually produces a group specific combination of services that meet the optimal non-inferior constraint condition. Users can choose the most satisfactory service combination according to need, Other services combination will as alternatives for using at the time of the accident.