الفهرس | Only 14 pages are availabe for public view |
Abstract The next generation of computing infrastructure is moving toward the cloud, and numerous cloud suppliers offer a wide range of cloud services. It can be very difficult to determine which cloud services are suitable for a certain application. In addition to the computing requirements, addressing this difficulty entails balancing a number of elements, including corporate demands, technology, policies, and preferences. Numerous cloud services have been developed as a result of the rapid growth of cloud computing. Any firm has the choice to adopt cloud services if it wants to maximize flexibility and respond quickly to market demands. It is a very major challenge for enterprises to choose the suitable cloud services that may satisfy their requirements because of the diversity of cloud service providers. As a result, choosing a cloud service provider can be seen as a form of decision analysis problem involving several stakeholders. Cloud service selection helps users choose the most suitable cloud services for their needs and reduces losses brought on by poor service selection. Therefore, the selection process of cloud services can be considered as a type of multi-criteria decision analysis problems. This thesis introduces a novel framework that can be used for selecting the most suitable provider in the case of missing values in the evaluation of alternatives based on a neutrosophic multi-criteria decision analysis (NMCDA) approach for assessing the quality of cloud services. The framework is composed of two steps:In first step, The Modified Generative Adversarial Network (M-GAN) is used in the framework to impute missing data. GANs are generative models based on the deep learning framework to generate artificial data. It provides a way to learn deep representations without extensively annotated training data. GANs comprise a generator and a discriminator, both trained under the adversarial learning idea. The goal of GANs is to estimate the potential distribution of real data samples and generate new samples from that distribution. The modified version of GAN has achieved an accuracy of nearly 94%. In second step, the multi-criteria decision-making neutrosophic algorithm that make evaluation of the alternatives for choosing the best provider in accordance with various eight criteria (Availability, Throughput, Successibility, Reliability, Latency, Response time, Response Time of Customer Services, and Cost). Multi Criteria Decision Making (MCDM) provides strong decision making in domains where selection of best alternative is highly complex. According to the experiments done in the thesis, the Novel framework has achieved success in choosing the most suitable cloud provider. |