Abstract
For the issue that users are often sensitive to the QoS (Quality of Service) of mobile applications, cloudlet has emerged as a novel paradigm which provides closer computing resources to users to preserve the QoS of mobile applications. By scheduling complex computing tasks to cloudlets, the energy consumption of mobile devices and the transmission latency of tasks are enabled to decrease. However, as the number of mobile applications access to cloudlets increases, preserving the QoS of mobile applications poses a challenge while offloading tasks to cloudlets. In view of this challenge, a QoS-aware work-flow scheduling method (MWSM) is designed in this paper. Technically, we model each mobile application to be scheduled as a workflow and analyze the workflow scheduling problem. Then, we formulate the QoS-aware workflow scheduling problem as a multi-objective optimization problem. Afterwards, NSGA-III (Non-dominated Sorting Genetic Algorithm III) is adopted to minimize energy consumption, transmission latency and operation cost of workflows. Furthermore, ELECTRE is employed to select the most optimal scheduling strategy. Finally, experimental evaluations are conducted to demonstrate the efficiency and potential of our proposed scheduling method.