Task Computing Maximization in MEC-enabled Cell-free Networks

Cell-free massive multiple-input multiple-output (CF-mMIMO) networks and mobile edge computing (MEC) are the two key enablers of next-generation wireless networks, offering enhanced connectivity, distributed and low-latency computation at the network edge. However, efficient task distribution and resource allocation are the critical challenges in MEC-enabled CF-mMIMO networks, directly impacting system performance. To address these challenges, we investigate the computing data maximization problem by optimizing the task distribution coefficients and time allocation for transmission and computing. In this work, we consider a MEC-enabled CF-mMIMO network with numerous distributed access points (APs), each of them is integrated with a MEC server, a few users (UEs) requiring to compute a large amount of data, and a central processing unit (CPU) coordinating the APs through the fronthaul. We propose two novel algorithms as the joint optimization of time allocation and task distribution coefficients (J-TATD) and load-aware weighted large-scale fading-based task distribution (LWL-TD) to maximize the computing data. The J-TATD algorithm jointly optimizes the time allocation for transmission and computing, and the task distribution coefficients, while the LWL-TD algorithm employs a heuristic approach and decouples the problem into two subproblems: i) determining task distribution coefficients; and ii) optimizing time allocation for transmission and computing using determined task distribution coefficients to maximize the total computing data. Performance evaluations demonstrate the effectiveness of the proposed J-TATD and LWL-TD algorithms including the impact of task distribution by implementing an equal task distribution (ETD) for CF-mMIMO networks compared to conventional co-located mMIMO and small-cell networks.