Abstract
Mobile edge computing (MEC) is becoming a promising paradigm to provide computing services for smart mobile devices (SMDs) via offloading computation-intensive tasks to MEC servers deployed at the network edge. In this paper, in order to further improve the accessing capacity of MEC systems and minimize all users' computation overhead, taking advantage of the superior spectral efficiency of Non-Orthogonal Multiple Access (NOMA) technology, we introduce NOMA into MEC systems and investigate a multi-user computation offloading problem through jointly optimizing offloading decisions, communication and computation resources allocation. To tackle the formulated mixed integer nonlinear programming (MINLP) problem which is NP-hard, we iteratively update either the resource allocation or the offloading decision via fixing the other solution and efficiently solve it in polynomial time. Specifically, given a fixed offloading decision, the sub-channel assignment problem is solved via applying a many-to-one matching model with peer effects, the transmission power of SMDs is optimized by combing sequential convex programming and parametric convex programming, and the computation resources allocation is addressed by convex optimization. Furthermore, the results of resource allocation are applied to guide the offloading decision. Extensive simulations show that our proposed algorithm performs closely to the optimal solution, and compared with existing solutions, our algorithm can efficiently improve the accessing capacity of MEC systems and reduce the total computation overhead of all users.
Highlights
In the fifth-generation (5G) era, smart mobile devices (SMDs) are taken as the dominant devices to support novel mobile applications such as virtual reality, face recognition, natural language processing, and interactive gaming
3) We further show that the joint allocation of communication and computation resources (JACCR) problem can be decomposed into two independent sub-problems: (i) communication resources allocation and (ii) computation resources allocation
As shown in Algorithm 5, the proposed joint offloading decision and resource allocation algorithm (JODRAA) consists of two phases: inner-loop and outer-loop, which deal with resource allocation and the offloading decision, respectively
Summary
In the fifth-generation (5G) era, smart mobile devices (SMDs) are taken as the dominant devices to support novel mobile applications such as virtual reality, face recognition, natural language processing, and interactive gaming. Combined with the results of computation resources allocation, it further affects the offloading decision and makes the computation offloading problem more challenging In this case, the main contributions of this paper are summarized as follows: 1) To improve the performance of MEC systems, we apply NOMA technology to transmit the offloading tasks and investigate the computation offloading mechanism. 2) To cope with the formulated NP-hard problem, through analyzing the underlying relationship between communication-computation resources allocation and the offloading decision, we design an alternating iteration algorithm to solve it in polynomial time. VOLUME 7, 2019 for a multi-user NOMA-MEC system via jointly optimizing the binary offloading decision, sub-channel assignment, the transmission power of SMDs, and computation resources allocation.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
