NOMA-Aided Mobile Edge Computing via User Cooperation

Yu‐Wen Huang, Yuan Liu, Fangjiong ChenIEEE Trans. Communications 2019

TL;DR

This paper proposes a novel mobile edge computing (MEC) framework that leverages Non-Orthogonal Multiple Access (NOMA) and user cooperation to optimize energy consumption and maximize offloaded data. By allowing users to offload tasks to both a helper device and the MEC access point (AP) simultaneously, the system achieves significant performance gains over traditional approaches.

What Problem Does This Paper Address?

The increasing demand for data-intensive mobile applications like augmented reality (AR) and virtual reality (VR) puts a strain on the limited computational resources and battery life of mobile devices. While Mobile Edge Computing (MEC) offers a solution by bringing computation closer to users, the limited capacity of MEC access points (APs) can become a bottleneck. To address this, the paper explores cooperative computing, where mobile devices can utilize each other's resources. However, efficiently managing communication and computation resources in such a cooperative environment, especially with the growing number of devices, is a significant challenge. This paper specifically tackles the problem of optimizing resource allocation in a three-node MEC system (user, helper, AP) using NOMA to enable efficient cooperative computing.

Methods

The core of the proposed solution lies in a NOMA-aided cooperative computing scheme. The user simultaneously offloads its computation task to both a helper device and the AP using NOMA in the first time slot. In the second time slot, the helper can then process its own task and offload a portion of the user's task to the AP. The paper formulates two optimization problems:

  1. Energy Consumption Minimization: This problem aims to minimize the total energy consumed by the user and helper by jointly optimizing task partitioning, transmit power, and offloading time. The authors transform the non-convex problem into a convex one and develop an efficient algorithm to find the optimal solution.
  2. Offloading Data Maximization: This problem focuses on maximizing the total amount of data offloaded by the user and helper. Similar to the first problem, it's a non-convex problem that is simplified and solved optimally.

Key techniques include:

  • NOMA: Enables simultaneous transmission to multiple users on the same resource block, improving spectral efficiency.
  • Cooperative Computing: Leverages the computational resources of the helper device.
  • Partial Offloading: Allows tasks to be split between local computation and offloading.
  • Convex Optimization: Used to solve the formulated non-convex problems by transforming them into convex equivalents.
  • Lagrangian Duality and Ellipsoid Method: Employed for solving the optimization problems.

Datasets

Not discussed in the paper. The evaluation is based on simulated scenarios with parameters defined in Table I.

Key Insights and Ideas

  • NOMA + Cooperation is a Powerful Combo: Combining NOMA with user cooperation significantly outperforms traditional TDMA and NOMA-without-cooperation schemes in both energy efficiency and offloading capacity.
  • Helper's Role is Crucial: The helper device acts as a vital intermediary, not only processing its own tasks but also aiding the user in offloading to the AP, thereby reducing the user's latency and energy burden.
  • Task Partitioning Matters: The optimal split of tasks between local computation and offloading, as well as between offloading to the helper and the AP, is critical for achieving performance gains.
  • Distance and SNR Influence Strategy: The optimal resource allocation strategy adapts based on the distances between devices and their signal-to-noise ratios (SNRs). For instance, in high SNR scenarios, offloading directly to the AP becomes more favorable.
  • Win-Win Scenarios: The proposed scheme can create win-win situations where both the user and helper benefit from cooperation, especially when latency constraints are tight.

Key Results

  • Energy Savings: The proposed NOMA-aided cooperative scheme achieves approximately 24% energy consumption reduction on average compared to benchmark schemes without cooperation.
  • Performance Improvement: The scheme provides a performance gap over benchmark schemes, demonstrating its effectiveness.
  • Latency Impact: As the latency constraint (T) increases, energy consumption decreases for all schemes, with the proposed scheme showing a smoother performance curve.
  • Offloading Data: Offloading data increases linearly with the latency constraint, and the proposed scheme consistently achieves the best offloading performance.
  • Power Allocation: The power allocation proportion between the user and helper adapts based on factors like user transmit power and distance, favoring offloading to the AP when it's resource-rich.

Limitations and Future Work

  • Helper's Task Dependency: The current model assumes the helper can process its own task and assist the user. Future work could explore scenarios where the helper's primary role is solely to assist the user.
  • Global CSI Assumption: The paper assumes global Channel State Information (CSI) is available, which might not always be practical in real-world scenarios. Future research could investigate the impact of imperfect CSI.
  • Simplified Model: The paper simplifies the system by assuming single-antenna half-duplex nodes and negligible downloading time. Extending the model to multi-antenna systems or considering download latency could provide a more comprehensive analysis.
  • Incentive Mechanisms: While the paper focuses on resource optimization, future work could incorporate incentive mechanisms to encourage user participation in cooperative computing.

Summary updated

Abstract

Exploiting the idle computation resources of mobile devices in mobile edge computing (MEC) system can achieve both channel diversity and computing diversity as mobile devices can offload their computation tasks to nearby mobile devices in addition to MEC server embedded access point (AP). In this paper, we propose a non-orthogonal multiple-access (NOMA)-aided cooperative computing scheme in a basic three-node MEC system consisting of a user, a helper, and an AP. In particular, we assume that the user can simultaneously offload data to the helper and the AP using NOMA, while the helper can locally compute data and offload data to the AP at the same time. We study two optimization problems, energy consumption minimization and offloading data maximization, by joint communication and computation resource allocation of the user and helper. We find the optimal solutions for the two non-convex problems by some proper mathematical methods. Simulation results are presented to demonstrate the effectiveness of the proposed schemes. Some useful insights are provided for practical designs.
BibTeX
@article{huang2019noma,
  author = {Yu‐Wen Huang and Yuan Liu and Fangjiong Chen},
  title = {{NOMA-Aided Mobile Edge Computing via User Cooperation}},
  journal = {IEEE Trans. Communications},
  year = {2019},
}
Figure