[Paper Review] Resource Allocation in Intelligent Reflecting Surface Assisted NOMA Systems

Jiakuo Zuo, Yuanwei Liu+2 more|arXiv (Cornell University)|Feb 5, 2020

Advanced Wireless Communication Technologies34 references17 citations

TL;DR

This paper proposes a three-step resource allocation algorithm for intelligent reflecting surface (IRS)-assisted non-orthogonal multiple access (NOMA) systems to maximize system throughput by jointly optimizing channel assignment, decoding order, power allocation, and reflection coefficients. The method achieves significant spectral efficiency gains over conventional NOMA and IRS-OMA systems, with performance enhanced by optimal IRS deployment.

ABSTRACT

This paper investigates the downlink communications of intelligent reflecting surface (IRS) assisted non-orthogonal multiple access (NOMA) systems. To maximize the system throughput, we formulate a joint optimization problem over the channel assignment, decoding order of NOMA users, power allocation, and reflection coefficients. The formulated problem is proved to be NP-hard. To tackle this problem, a three-step novel resource allocation algorithm is proposed. Firstly, the channel assignment problem is solved by a many-to-one matching algorithm. Secondly, by considering the IRS reflection coefficients design, a low-complexity decoding order optimization algorithm is proposed. Thirdly, given a channel assignment and decoding order, a joint optimization algorithm is proposed for solving the joint power allocation and reflection coefficient design problem. Numerical results illustrate that: i) with the aid of IRS, the proposed IRS-NOMA system outperforms the conventional NOMA system without the IRS in terms of system throughput; ii) the proposed IRS-NOMA system achieves higher system throughput than the IRS assisted orthogonal multiple access (IRS-OMA) systems; iii) simulation results show that the performance gains of the IRS-NOMA and the IRS-OMA systems can be enhanced via carefully choosing the location of the IRS.

Motivation & Objective

• Address the challenge of maximizing system throughput in IRS-assisted NOMA downlink systems with joint optimization of multiple resource allocation variables.
• Overcome the NP-hard complexity of joint channel assignment, decoding order, power allocation, and reflection coefficient design in IRS-NOMA systems.
• Develop a low-complexity, near-optimal solution that outperforms conventional NOMA and IRS-OMA systems in spectral efficiency and coverage.
• Investigate the impact of IRS location on system performance to guide practical deployment.

Proposed method

• Formulates a joint optimization problem over channel assignment, decoding order, power allocation, and IRS reflection coefficients, proven to be NP-hard.
• Solves channel assignment via a many-to-one matching algorithm based on user and channel utilities, ensuring stability and convergence.
• Proposes a low-complexity decoding order optimization algorithm that leverages user channel gains and IRS-induced phase shifts to avoid exhaustive search.
• Applies alternating optimization and successive convex approximation (SCA) to jointly solve power allocation and reflection coefficient design under power and phase constraints.
• Uses a swap-operation-based refinement in the matching algorithm to improve utility and ensure convergence.
• Employs a three-stage sequential optimization: (1) matching-based channel assignment, (2) decoding order selection, (3) joint power and phase optimization.

Experimental results

Research questions

• RQ1 How can joint optimization of channel assignment, decoding order, power allocation, and IRS reflection coefficients maximize system throughput in IRS-NOMA systems?
• RQ2 What is the performance gain of IRS-NOMA over conventional NOMA and IRS-OMA systems in terms of spectral efficiency?
• RQ3 How does the location of the IRS affect the system throughput in IRS-NOMA deployments?
• RQ4 Can a low-complexity algorithm achieve near-optimal performance compared to exhaustive search methods?
• RQ5 What is the impact of IRS-induced channel enhancement on user fairness and spectral efficiency in NOMA systems?

Key findings

• The proposed IRS-NOMA system achieves a system throughput gain of 4.1 bit/s/Hz over conventional NOMA without IRS at an IRS location of 10 m.
• At an IRS location of 45 m, the system throughput gain increases to 6.1 bit/s/Hz due to stronger IRS-user links.
• The proposed three-step algorithm achieves near-optimal performance with significantly lower complexity than exhaustive search, particularly in channel assignment.
• The decoding order optimization algorithm achieves comparable performance to exhaustive search but with much lower computational cost.
• IRS-aided NOMA outperforms IRS-OMA systems in spectral efficiency, demonstrating the synergy between NOMA’s multiplexing gain and IRS’s channel enhancement.
• System performance is highly sensitive to IRS deployment; optimal placement near users maximizes the combined channel gain and throughput.