Mobile edge computing (MEC) has restricted computation resource and limited bandwidth resource when serving multiple users simultaneously. In general, a user’s task can not be always parallelized. In this paper, we study a joint computation and bandwidth resource allocation problem, considering a generalized task computing model. The computation capacity is modeled as a non-decreasing concave function about allocated computation resource, and the joint computation and bandwidth resources allocation problem is solved by cvx opt. framework. Simulation results show that our proposed scheme is greatly superior to the serial processing benchmark scheme for Non-linear model. With non-linear model, higher computational multiplexing gain can be achieved, however, for linear model, there is no gain that can be obtained from the proposed scheme.
View full abstract