Multi-User Task Offloading Strategy in RIS-Aided Multi-AP Mobile Edge Computing Networks

• Published in: IEEE access Vol. 13; pp. 118759 - 118770
• Main Authors: Zhou, Wen, Miao, Ling, Deng, Dan, Hua, Min, Xiang, Dan, Li, Chunguo
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Autonomous aerial vehicles; branch and bound (BnB); Computation offloading; Constraints; Costs; Delay effects; Edge computing; Energy consumption; Lower bounds; Mixed integer; Mobile computing; Mobile edge computing (MEC); Multi-access edge computing; Network latency; Nonlinear programming; reconfigurable intelligent surface (RIS); Reconfigurable intelligent surfaces; Reflection coefficient; Servers; Vectors; Wireless communication; Wireless communications
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2025.3585551

The incorporation of RIS into the mobile edge computing (MEC) network can improve the wireless communication environment and enhance the task-offloading capability of the network. In this paper, we consider a reconfigurable intelligent surface (RIS)-aided edge computing-enabled multiuser networks with multiple access points (APs) and investigate the computation offloading strategy of user equipments (UEs). The user association matrix and the RIS reflecting coefficients are jointly optimized to minimize the energy consumption of UEs, subject to the latency constraint and other ones. We formulate it as a mixed integer nonlinear programming (MINLP). To solve the MINLP, we propose a branch and bound (BnB) based method, in which two important links are tackled, i.e., finding the MINLP's lower bound and updating the incumbent solution. For the first link, the objective function and the latency constraint are relaxed to render the problem more tractable. For the second link, a lemma is first introduced to address the nonconvex phase constraint by establishing the equivalence between the original problem and the transformed problem. Subsequently, we propose an algorithm based on successive convex approximation (SCA), constructing a sequence of convex subproblems. To enhance the feasibility of the algorithm, a subgradient method is further employed to solve each subproblem. The convergence and complexity of the proposed method are analyzed and its effectiveness is demonstrated through simulation results.