A Multi-Objective Clustering Evolutionary Algorithm for Multi-Workflow Computation Offloading in Mobile Edge Computing

Lei Pan,Xiao Liu,Zhaohong Jia,Jia Xu,Xuejun Li

DOI: https://doi.org/10.1109/tcc.2021.3132175

IF: 5.697

2023-01-01

IEEE Transactions on Cloud Computing

Abstract:To cope with the rapid development of the Internet of Things (IoT) and the increasing demand for real-time services, mobile edge computing (MEC) has become a promising solution which extends centralised cloud computing, to provision computing resources, storage and network services closer to the mobile device from the network edge. While computation offloading is a key feature in MEC to enable real-time services, offloading workflow tasks in MEC is an NP-hard problem. Typically, the problem of multi-workflow offloading with multi-objective optimization is still an open and challenging issue. Therefore, this article proposes a multi-objective clustering evolutionary algorithm called MCEA to minimize the cost and energy consumption of multi-workflow execution under the deadline constraint. First, the sub-deadline constraint is added during initialization to generate more initial solutions that satisfy the deadline constraint. Then an adaptive clustering method is adopted to guide individuals to find a suitable mate during crossover operation. Finally, the probabilities of crossover and mutation are dynamically adjusted based on the historical information to control the evolution direction and convergence speed of algorithm. Comprehensive experiments are carried out for complex workflow applications on FogWorkflowSim, which demonstrate that MCEA can achieve better performance than four representative algorithms in three evaluation metrics.