QoE-Aware User Allocation in NOMA-Enabled MEC Systems: A Distributed Game-Theoretical Approach

In vehicular networks, mobile edge computing (MEC) allows service providers to allocate vehicle terminals (VTs) to edge servers, providing an effective paradigm for low latency computing services. In resources-limited edge computing scenarios, user allocation needs to achieve the goals of minimizing deployment costs and maximizing coverage density while ensuring service quality. However, the differentiated configuration of user power in non-orthogonal multiple access (NOMA) technology can cause intra-cell and inter-cell interference within the same frequency band, resulting in a significant decrease in transmission rate and a reduction in Quality of Experience (QoE). In this paper, we study the QoE-aware user and power allocation (QUPA) problem with multiple users in a NOMA-enabled MEC system aimed at maximizing users' QoE. We formulate this problem as a QUPA game and theoretically analyze its properties. To achieve the Nash equilibrium for each user, we propose a distributed game-theoretical user and power allocation (GUPA) algorithm that jointly optimizes server selection, channel selection, and transmission power allocation. Meanwhile, we theoretically analyze the convergence and performance by price of anarchy (PoA) of our GUPA algorithm. The experimental results show that the proposed GUPA algorithm can effectively reduce costs and improve user experience in this system.