Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning
In this paper, a system utilizing an active intelligent reflecting surface (IRS) to enhance the performance of wireless communication network is modeled, which has the ability to adjust power between base station (BS) and active IRS. We aim to maximize the signal-to-noise ratio (SNR) of the user by...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Spolecnost pro radioelektronicke inzenyrstvi
2024-12-01
|
| Series: | Radioengineering |
| Subjects: | |
| Online Access: | https://www.radioeng.cz/fulltexts/2024/24_04_0571_0582.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850271481897943040 |
|---|---|
| author | Q. Cheng J. Bai X. Wang B. Shi W. Gao F. Shu |
| author_facet | Q. Cheng J. Bai X. Wang B. Shi W. Gao F. Shu |
| author_sort | Q. Cheng |
| collection | DOAJ |
| description | In this paper, a system utilizing an active intelligent reflecting surface (IRS) to enhance the performance of wireless communication network is modeled, which has the ability to adjust power between base station (BS) and active IRS. We aim to maximize the signal-to-noise ratio (SNR) of the user by jointly designing power allocation (PA) factor, active IRS phase shift matrix, and beamforming vector of BS, subject to a total power constraint. To tackle this non-convex problem, we solve this problem by alternately optimizing these variables. The PA factor is designed via polynomial regression method in machine learning. BS beamforming vector and IRS phase shift matrix are obtained by Dinkelbach's transform and successive convex approximation methods. Then, we maximize achievable rate (AR) and use closed-form fractional programming (CFFP) method to transform the original problem into an equivalent form. This problem is addressed by iteratively optimizing auxiliary variables, BS and IRS beamformings. Thus, two iterative PA methods are proposed accordingly, namely maximizing SNR based on PA factor (Max-SNR-PA) and maximizing AR based on CFFP (Max-AR-CFFP). The former has a better rate performance, while the latter has a lower computational complexity. Simulation results show that the proposed algorithms can effectively improve the rate performance compared to fixed PA strategies, only optimizing PA factor, aided by passive IRS, and without IRS. |
| format | Article |
| id | doaj-art-2724e2dbf0004884bd841b3f04f21483 |
| institution | OA Journals |
| issn | 1210-2512 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Spolecnost pro radioelektronicke inzenyrstvi |
| record_format | Article |
| series | Radioengineering |
| spelling | doaj-art-2724e2dbf0004884bd841b3f04f214832025-08-20T01:52:14ZengSpolecnost pro radioelektronicke inzenyrstviRadioengineering1210-25122024-12-01334571582Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine LearningQ. ChengJ. BaiX. WangB. ShiW. GaoF. ShuIn this paper, a system utilizing an active intelligent reflecting surface (IRS) to enhance the performance of wireless communication network is modeled, which has the ability to adjust power between base station (BS) and active IRS. We aim to maximize the signal-to-noise ratio (SNR) of the user by jointly designing power allocation (PA) factor, active IRS phase shift matrix, and beamforming vector of BS, subject to a total power constraint. To tackle this non-convex problem, we solve this problem by alternately optimizing these variables. The PA factor is designed via polynomial regression method in machine learning. BS beamforming vector and IRS phase shift matrix are obtained by Dinkelbach's transform and successive convex approximation methods. Then, we maximize achievable rate (AR) and use closed-form fractional programming (CFFP) method to transform the original problem into an equivalent form. This problem is addressed by iteratively optimizing auxiliary variables, BS and IRS beamformings. Thus, two iterative PA methods are proposed accordingly, namely maximizing SNR based on PA factor (Max-SNR-PA) and maximizing AR based on CFFP (Max-AR-CFFP). The former has a better rate performance, while the latter has a lower computational complexity. Simulation results show that the proposed algorithms can effectively improve the rate performance compared to fixed PA strategies, only optimizing PA factor, aided by passive IRS, and without IRS.https://www.radioeng.cz/fulltexts/2024/24_04_0571_0582.pdfactive intelligent reflecting surfaceachievable ratepower allocationclosed-form fractional programming |
| spellingShingle | Q. Cheng J. Bai X. Wang B. Shi W. Gao F. Shu Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning Radioengineering active intelligent reflecting surface achievable rate power allocation closed-form fractional programming |
| title | Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning |
| title_full | Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning |
| title_fullStr | Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning |
| title_full_unstemmed | Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning |
| title_short | Two Power Allocation and Beamforming Strategies for Active IRS-aided Wireless Network via Machine Learning |
| title_sort | two power allocation and beamforming strategies for active irs aided wireless network via machine learning |
| topic | active intelligent reflecting surface achievable rate power allocation closed-form fractional programming |
| url | https://www.radioeng.cz/fulltexts/2024/24_04_0571_0582.pdf |
| work_keys_str_mv | AT qcheng twopowerallocationandbeamformingstrategiesforactiveirsaidedwirelessnetworkviamachinelearning AT jbai twopowerallocationandbeamformingstrategiesforactiveirsaidedwirelessnetworkviamachinelearning AT xwang twopowerallocationandbeamformingstrategiesforactiveirsaidedwirelessnetworkviamachinelearning AT bshi twopowerallocationandbeamformingstrategiesforactiveirsaidedwirelessnetworkviamachinelearning AT wgao twopowerallocationandbeamformingstrategiesforactiveirsaidedwirelessnetworkviamachinelearning AT fshu twopowerallocationandbeamformingstrategiesforactiveirsaidedwirelessnetworkviamachinelearning |