TY - JOUR
T1 - Hyper-angle exploitative searching for enabling multi-objective optimization of fog computing
AU - Abdali, Taj Aldeen Naser
AU - Hassan, Rosilah
AU - Aman, Azana Hafizah Mohd
AU - Nguyen, Quang Ngoc
AU - Al-Khaleefa, Ahmed Salih
N1 - Funding Information:
Funding: This paper is supported under grant Fundamental Research Grant Scheme FRGS/1/2018/TK04/UKM/02/17 and Dana Impak Perdana UKM DIP-2018-040.
Funding Information:
Acknowledgments: The authors would like to acknowledge the support provided by the Network and Communication Technology (NCT) Research Groups, FTSM, UKM in providing facilities throughout this paper. The authors would also like to thank the Editor and the anonymous reviewers for their valuable comments and suggestions.
Publisher Copyright:
© 2021 by the authors.
PY - 2021/1/2
Y1 - 2021/1/2
N2 - Fog computing is an emerging technology. It has the potential of enabling various wireless networks to offer computational services based on certain requirements given by the user. Typically, the users give their computing tasks to the network manager that has the responsibility of allocating needed fog nodes optimally for conducting the computation effectively. The optimal allocation of nodes with respect to various metrics is essential for fast execution and stable, energy-efficient, bal-anced, and cost-effective allocation. This article aims to optimize multiple objectives using fog computing by developing multi-objective optimization with high exploitive searching. The developed algorithm is an evolutionary genetic type designated as Hyper Angle Exploitative Searching (HAES). It uses hyper angle along with crowding distance for prioritizing solutions within the same rank and selecting the highest priority solutions. The approach was evaluated on multi-objective mathematical problems and its superiority was revealed by comparing its performance with benchmark approaches. A framework of multi-criteria optimization for fog computing was proposed, the Fog Computing Closed Loop Model (FCCL). Results have shown that HAES outperforms other relevant benchmarks in terms of non-domination and optimality metrics with over 70% confidence of the t-test for rejecting the null-hypothesis of non-superiority in terms of the domination metric set coverage.
AB - Fog computing is an emerging technology. It has the potential of enabling various wireless networks to offer computational services based on certain requirements given by the user. Typically, the users give their computing tasks to the network manager that has the responsibility of allocating needed fog nodes optimally for conducting the computation effectively. The optimal allocation of nodes with respect to various metrics is essential for fast execution and stable, energy-efficient, bal-anced, and cost-effective allocation. This article aims to optimize multiple objectives using fog computing by developing multi-objective optimization with high exploitive searching. The developed algorithm is an evolutionary genetic type designated as Hyper Angle Exploitative Searching (HAES). It uses hyper angle along with crowding distance for prioritizing solutions within the same rank and selecting the highest priority solutions. The approach was evaluated on multi-objective mathematical problems and its superiority was revealed by comparing its performance with benchmark approaches. A framework of multi-criteria optimization for fog computing was proposed, the Fog Computing Closed Loop Model (FCCL). Results have shown that HAES outperforms other relevant benchmarks in terms of non-domination and optimality metrics with over 70% confidence of the t-test for rejecting the null-hypothesis of non-superiority in terms of the domination metric set coverage.
KW - Crowding distance
KW - Evolutionary genetics
KW - Fog computing
KW - Hy-per-angle
KW - Multi-objective optimization
KW - Task allocation
UR - http://www.scopus.com/inward/record.url?scp=85099348505&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85099348505&partnerID=8YFLogxK
U2 - 10.3390/s21020558
DO - 10.3390/s21020558
M3 - Article
AN - SCOPUS:85099348505
VL - 21
SP - 1
EP - 28
JO - Sensors (Switzerland)
JF - Sensors (Switzerland)
SN - 1424-3210
IS - 2
M1 - 558
ER -