Joint Task Offloading and Computation in Cooperative Multicarrier Relaying-Based Mobile-Edge Computing Systems

This article studies a mobile-edge computing (MEC) system, where an access point (AP) and a relay node serve a user terminal over multicarrier subchannels. In the MEC system, the relay can assist not only task offloading but also task computation. Aiming at minimizing total energy consumption at the user terminal and the relay, the resource allocations, such as subcarrier allocation, power allocation, task partition, and offloading time and computation time allocation, are to be optimized, subject to a given task computation delay constraint. To achieve this goal, a novel cooperative MEC protocol is designed, where multicarrier subchannels are utilized for parallel task offloading by integrating the rateless coding technique. Then, under the newly designed protocol, the resource allocation optimization problem is formulated as a mixed-integer programming (MIP) problem that is challenging to solve. To tackle this MIP problem, continuous relaxation and algebraic transformation techniques are applied to transform it into a convex problem in order to reveal the lower bound of energy consumption performance. After that, by equivalently rewriting the integer subcarrier allocation constraint in the original optimization problem as the intersection of a convex set and a d.c. (difference of two convex sets) set, the problem is solved by the successive convex approximation to achieve a practical and efficient resource allocation scheme. Simulation results show that the proposed jointly cooperative task offloading and computation scheme can significantly reduce the energy consumption as compared to the baseline schemes, where the relay only assists the task offloading or task computation.

Automated summary

This article studies a mobile-edge computing (MEC) system and designs a novel cooperative MEC protocol to minimize total energy consumption at the user terminal and the relay through optimized resource allocations. The simulation results show that the proposed scheme can significantly reduce energy consumption compared to baseline schemes.