| Abstract |
With the proliferation of mobile applications and the explosive growth of mobile devices, huge computing demands are generated, leading to more computational delays. The occur of mobile edge computing (MEC) pushes calculations to the edge of the network and then realize compute-intensive tasks on resource-constrained devices, in which task allocation is more difficult. In this paper, we design a particle swarm optimization (PSO) and game theoretic based task allocation for MEC. Firstly, in order to ensure nodes in the same group are closer, the maximizing minimum distance clustering algorithm is designed to generate the structure of the parallel group. Secondly, we propose a multi-task assignment model based on Nash equilibrium, and design the strategies of each task and the utility function. Then we use the PSO to find the Nash equilibrium point, minimizing the all tasks execution time and saving the energy cost and find the tasks that need to be offloaded to the group (the group is made up of base stations). Moreover, we use the priority setting algorithm to sort tasks and then upload tasks to the group in a certain order, thereby confirming the order of tasks uploaded on the device, which jointly considers the calculation time in base station and mobile device and transmission time. In addition, a task migration algorithm between the groups is proposed for congested groups, which is based on the group migration time and waiting time. Simulation results demonstrate the effectiveness of the PSOGT and it can effectively reduce delay for MEC. © 2019 IEEE. |
