Task-Oriented Delay-Aware Multi-Tier Computing in Cell-free Massive MIMO Systems

Multi-tier computing can enhance the task computation by multi-tier nodes. In this paper, we propose a cell-free massive multiple-input multiple-output (MIMO) aided system deploying nodes to improve performance. At first, investigate computational latency and total energy consumption for computation, regarded as cost. Then, formulate cost minimization problem design bandwidth allocation allocation, while considering realistic heterogenous delay requirements of tasks. Due binary variable, formulated optimization is non-convex. Therefore, solve decoupling original into subproblems. As convex problem, first determine given strategy, followed conceiving traditional strategy obtain solution. Based on asymptotic property received signal-to-interference-plus-noise ratio (SINR) under MIMO setting solution, dual subproblem relaxing constraint with Lagrange partial relaxation requirements. last, simulation results are provided demonstrate that our proposed offloading scheme performs better than benchmark schemes, where minimum-cost optimal heterogeneous tasks may be controlled SINR in MIMO-aided systems.