Joint computation offloading and resource allocation for NOMA-based multi-access mobile edge computing systems

Due to explosive growth of computation-intensive applications, multi-access mobile edge computing (MEC) has been proposed as an efficient technology to relieve the burden on mobile users by offloading computation tasks to MEC servers. Meanwhile, nonorthogonal multiple access (NOMA) has been proposed to improve spectrum efficiency by allowing multiple users transmitting on the same time and spectrum resource. Thus, NOMA can be naturally adopted by the MEC systems to improve the offloading performance. In this paper, we consider a NOMA-based multi-access MEC system with multiple MEC servers and multiple users. It is assumed that each user has multiple computation tasks, where each task shall be offloaded to one MEC server for computing. NOMA is applied for uploading multiple tasks to MEC servers as well as downloading multiple task results. The problem of jointly optimizing the MEC server allocation, the transmit power allocation of the users, the transmit power allocation of the MEC servers, the computation resource allocation of the MEC servers, the time allocation and the channel allocation, for minimizing the overall delay of all tasks is investigated. We then propose a heuristic algorithm to solve the problem, where the computation resource allocation of the MEC servers and the time allocation are optimized by the bisection search method, the transmit power allocations are obtained in closed-form, and the MEC server allocation and the channel allocation are optimized by the swap matching algorithm. Extensive simulations demonstrate that the proposed algorithm significantly outperforms various benchmark algorithms.