[Paper Review] Efficient Resource Allocation for Relay-Assisted Computation Offloading in Mobile Edge Computing

Xihan Chen, Yunlong Cai+4 more|arXiv (Cornell University)|Sep 1, 2019

Advanced Wireless Communication Technologies25 references4 citations

TL;DR

This paper proposes a hybrid relaying (HR) scheme in relay-assisted computation offloading (RACO) for mobile edge computing, using two orthogonal frequency bands: amplify-and-forward for result sharing and decode-and-forward for task transfer. By jointly optimizing offloading ratio, bandwidth, processor speeds, and transmit power, the method minimizes weighted sum delay and energy consumption via a CCCP-based iterative algorithm, achieving superior performance over benchmarks with reduced complexity.

ABSTRACT

In this article, we consider the problem of relay assisted computation offloading (RACO), in which user A aims to share the results of computational tasks with another user B through wireless exchange over a relay platform equipped with mobile edge computing capabilities, referred to as a mobile edge relay server (MERS). To support the computation offloading, we propose a hybrid relaying (HR) approach employing two orthogonal frequency bands, where the amplify-and-forward scheme is used in one band to exchange computational results, while the decode-and-forward scheme is used in the other band to transfer the unprocessed tasks. The motivation behind the proposed HR scheme for RACO is to adapt the allocation of computing and communication resources both to dynamic user requirements and to diverse computational tasks. Within this framework, we seek to minimize the weighted sum of the execution delay and the energy consumption in the RACO system by jointly optimizing the computation offloading ratio, the bandwidth allocation, the processor speeds, as well as the transmit power levels of both user $A$ and the MERS, under practical constraints on the available computing and communication resources. The resultant problem is formulated as a non-differentiable and nonconvex optimization program with highly coupled constraints. By adopting a series of transformations and introducing auxiliary variables, we first convert this problem into a more tractable yet equivalent form. We then develop an efficient iterative algorithm for its solution based on the concave-convex procedure. By exploiting the special structure of this problem, we also propose a simplified algorithm based on the inexact block coordinate descent method, with reduced computational complexity. Finally, we present numerical results that illustrate the advantages of the proposed algorithms over state-of-the-art benchmark schemes.

Motivation & Objective

• Address the challenge of low-latency, energy-efficient computation offloading in mobile edge computing under dynamic user and task demands.
• Enable efficient sharing of computational results between two users via a mobile edge relay server (MERS) in a relay-assisted setup.
• Minimize the weighted sum of execution delay and energy consumption under practical constraints on computing and communication resources.
• Jointly optimize computation offloading ratio, bandwidth allocation, processor speeds, and transmit power for both users and the MERS.
• Develop low-complexity algorithms that effectively solve the non-differentiable, nonconvex optimization problem with highly coupled constraints.

Proposed method

• Propose a hybrid relaying (HR) scheme using two orthogonal frequency bands: amplify-and-forward for result exchange and decode-and-forward for unprocessed task transfer.
• Formulate the resource allocation problem as a non-differentiable, nonconvex optimization program with coupled constraints on computing and communication resources.
• Transform the original problem into an equivalent, more tractable form using variable substitutions and auxiliary variables.
• Develop an iterative algorithm based on the concave-convex procedure (CCCP) to solve the reformulated problem with guaranteed convergence.
• Propose a simplified algorithm using the inexact block coordinate descent method to reduce computational complexity while maintaining performance.
• Solve the dual problem via bisection method with closed-form updates for power allocation, leveraging first-order optimality conditions and projection onto feasible sets.

Experimental results

Research questions

• RQ1 How can resource allocation be jointly optimized in a relay-assisted computation offloading system to minimize both delay and energy consumption?
• RQ2 What is the optimal trade-off between computation offloading and communication relaying in a two-user MEC scenario with a mobile edge relay server?
• RQ3 How can the non-differentiable and nonconvex nature of the joint optimization problem be effectively addressed with low-complexity algorithms?
• RQ4 What performance gains can be achieved by using a hybrid relaying scheme over conventional relaying methods in RACO systems?
• RQ5 How do dynamic user requirements and diverse computational tasks influence the design of adaptive resource allocation strategies in mobile edge networks?

Key findings

• The proposed CCCP-based iterative algorithm achieves convergence and significantly outperforms state-of-the-art benchmark schemes in minimizing the weighted sum of delay and energy consumption.
• The simplified inexact block coordinate descent algorithm reduces computational complexity while maintaining near-optimal performance, making it suitable for real-time deployment.
• Numerical results demonstrate that the hybrid relaying scheme effectively balances task offloading and result sharing, leading to lower end-to-end delay and reduced energy usage.
• The bisection-based dual decomposition method enables efficient solution of the dual problem with closed-form power updates, ensuring fast convergence.
• The algorithmic design successfully handles the non-differentiable and nonconvex nature of the optimization problem through successive convex approximations.
• The proposed system achieves a notable reduction in weighted sum cost compared to conventional decode-and-forward or amplify-and-forward only schemes, especially under high load conditions.