cs 20(18): e2

Research Article

An Intelligent Machine Learning and Self Adaptive Resource Allocation Framework for Cloud Computing Environment

Download67 downloads
  • @ARTICLE{10.4108/eai.13-7-2018.165501,
        author={Md. Shahidul Hasan and Balamurugan E and Md. Shawkat Akbar Almamun and Sangeetha K},
        title={An Intelligent Machine Learning and Self Adaptive Resource Allocation Framework for Cloud Computing Environment},
        journal={EAI Endorsed Transactions on Cloud Systems},
        volume={6},
        number={18},
        publisher={EAI},
        journal_a={CS},
        year={2020},
        month={6},
        keywords={Cloud Computing, Resource Allocation, QoS prediction model, Improved Bat Algorithm (IBA), Energy Efficient Model (EEM), Modified Clonal Selection Algorithm (MCSA), Enhanced Recurrent Neural Network},
        doi={10.4108/eai.13-7-2018.165501}
    }
    
  • Md. Shahidul Hasan
    Balamurugan E
    Md. Shawkat Akbar Almamun
    Sangeetha K
    Year: 2020
    An Intelligent Machine Learning and Self Adaptive Resource Allocation Framework for Cloud Computing Environment
    CS
    EAI
    DOI: 10.4108/eai.13-7-2018.165501
Md. Shahidul Hasan1,*, Balamurugan E2, Md. Shawkat Akbar Almamun1, Sangeetha K2
  • 1: Research Scholar, Texila American University, Guyana
  • 2: University of Africa, Toru-Orua, Nigeria
*Contact email: rethinbs@gmail.com

Abstract

Resource allocation is one of the major concern in cloud computing model. When several problems exists in rendering a useful resource allocator. In this research, a self adaptive resource allocation frame work based on machine learning is proposed for modelling and analysing the problem of multi-dimensional cloud resource. This novel self-adaptive resource allocation architecture consists of three stages, QoS prediction model, Improved Bat Algorithm (IBA) and Energy Efficient Model (EEM). The first one, the QoS prediction model, which depends on the same scale of system’s past events data, can attain a comparable accuracy with regard to QoS prediction. Secondly, an Energy Efficient Model, which is based on Modified Clonal Selection Algorithm (MCSA) is introduced for minimizing the energy depletion. Thirdly, a runtime decision-making algorithm that depends on improved bat algorithm can rapidly decide on a suitable function for resource allocation.