Muhammad Ali
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Muhammad Ali has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Fazal Sajid
School of Electronic and Information Engineering, Changchun University of Science and Technology, Changchun, China
3 shared publications - Jen-Yi Pan
Department of Communications Engineering, National Chung Cheng University, Minxiong, Taiwan
3 shared publications - Tabinda Ashraf
Department of Communications Engineering, National Chung Cheng University, Minxiong, Taiwan
3 shared publications - Muhammad Iqbal
Department of Communications Engineering, National Chung Cheng University, Minxiong, Taiwan
2 shared publications - Muhammad Sami Akram
Department of Electrical Engineering, National University of Sciences and Technology (NUST), Islamabad, Pakistan
2 shared publications - Anna Maria Vegni
Department of Industrial, Electronics and Mechanical Engineering, Roma Tre University, Rome, Italy
1 shared publication
Latest Publications
- Open Access
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Computer Society
- Vol. 7, 1015-1025
Hybrid RIS Phase-Shift Design via Rao-Enhanced Whale Optimization for SE Maximization in Downlink MISO HetNets
In this paper, we address the problem of maximizing spectral efficiency (SE) for the downlink of a hybrid reconfigurable intelligent surface (RIS)-assisted multiple-input single-output (MISO) heterogeneous network (HetNet). We introduce a hybrid RIS architecture in which a small fraction of RIS elements operate as active amplifying elements, while the remaining elements operate as passive reflectors to assist the small base station (SBS)–user equipment (UE) link. We employ zero-forcing (ZF) beamforming at the SBS and then optimize the hybrid RIS phase shifts via a Rao-enhanced Whale Optimization Algorithm (WOA-Rao). We benchmark WOA-Rao against particle swarm optimization (PSO), grey wolf optimizer (GWO), and semidefinite relaxation (SDR) under realistic path-loss, Rician/Rayleigh fading, noise, and power constraints. Extensive simulations demonstrate that WOA-Rao converges rapidly to high-quality solutions and consistently outperforms PSO, GWO, and SDR in SE gains and convergence speed; SE further improves with more RIS elements and higher SBS power, and an effective RIS placement distance is identified to balance SBS–RIS and RIS–UE link quality. An active-element-ratio ablation study further shows that the selected $12\%$ active-element ratio provides a practical SE and energy-efficiency (EE) tradeoff. These results demonstrate the effectiveness of hybrid active–passive RIS designs combined with metaheuristic optimization for next-generation wireless HetNets.
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Computer Society
- Vol. 7, 1015-1025
Open AccessHybrid RIS Phase-Shift Design via Rao-Enhanced Whale Optimization for SE Maximization in Downlink MISO HetNets
In this paper, we address the problem of maximizing spectral efficiency (SE) for the downlink of a hybrid reconfigurable intelligent surface (RIS)-assisted multiple-input single-output (MISO) heterogeneous network (HetNet). We introduce a hybrid RIS architecture in which a small fraction of RIS elements operate as active amplifying elements, while the remaining elements operate as passive reflectors to assist the small base station (SBS)–user equipment (UE) link. We employ zero-forcing (ZF) beamforming at the SBS and then optimize the hybrid RIS phase shifts via a Rao-enhanced Whale Optimization Algorithm (WOA-Rao). We benchmark WOA-Rao against particle swarm optimization (PSO), grey wolf optimizer (GWO), and semidefinite relaxation (SDR) under realistic path-loss, Rician/Rayleigh fading, noise, and power constraints. Extensive simulations demonstrate that WOA-Rao converges rapidly to high-quality solutions and consistently outperforms PSO, GWO, and SDR in SE gains and convergence speed; SE further improves with more RIS elements and higher SBS power, and an effective RIS placement distance is identified to balance SBS–RIS and RIS–UE link quality. An active-element-ratio ablation study further shows that the selected $12\%$ active-element ratio provides a practical SE and energy-efficiency (EE) tradeoff. These results demonstrate the effectiveness of hybrid active–passive RIS designs combined with metaheuristic optimization for next-generation wireless HetNets.
- 7 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. PP (00189545) , 1-14
Soft Actor-Critic for Robust Optimization in RIS-Assisted MU-MISO Systems
Reconfigurable intelligent surfaces (RIS) have emerged as a promising technology for enhancing spectral and energy efficiency in next-generation wireless networks. However, the joint optimization of base station (BS) precoding and RIS reflection coefficients in multi-user multiple-input single-output (MU-MISO) systems is a highly non-convex problem with prohibitive computational complexity when addressed using conventional optimization techniques. To tackle this challenge, we propose a data-driven reinforcement learning framework based on the Soft Actor-Critic (SAC) algorithm for adaptive RIS-assisted MU-MISO optimization. The proposed approach exploits stochastic policy learning with entropy regularization, twin Q-networks, and automatic temperature adjustment to enhance exploration capability and training stability. To rigorously benchmark performance, we conduct a comprehensive comparative analysis against two widely adopted deterministic reinforcement learning baselines, namely Deep Deterministic Policy Gradient (DDPG) and Twin Delayed Deep Deterministic Policy Gradient (TD3), under varying transmit power levels, RIS sizes, and learning hyperparameters. Simulation results demonstrate that SAC consistently achieves superior long-term sum-rate performance, particularly in low-to-moderate signal-to-noise ratio (SNR) regimes and for larger RIS dimensions, while exhibiting greater robustness to hyperparameter tuning. In contrast, DDPG and TD3 often converge faster in early training stages but are more sensitive to parameter settings and prone to suboptimal asymptotic solutions. These results establish SAC as a robust and scalable solution for RIS-assisted MU-MISO systems, offering improved reliability and adaptability for future intelligent wireless networks.
- 7 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. PP (00189545) , 1-14
Soft Actor-Critic for Robust Optimization in RIS-Assisted MU-MISO Systems
Reconfigurable intelligent surfaces (RIS) have emerged as a promising technology for enhancing spectral and energy efficiency in next-generation wireless networks. However, the joint optimization of base station (BS) precoding and RIS reflection coefficients in multi-user multiple-input single-output (MU-MISO) systems is a highly non-convex problem with prohibitive computational complexity when addressed using conventional optimization techniques. To tackle this challenge, we propose a data-driven reinforcement learning framework based on the Soft Actor-Critic (SAC) algorithm for adaptive RIS-assisted MU-MISO optimization. The proposed approach exploits stochastic policy learning with entropy regularization, twin Q-networks, and automatic temperature adjustment to enhance exploration capability and training stability. To rigorously benchmark performance, we conduct a comprehensive comparative analysis against two widely adopted deterministic reinforcement learning baselines, namely Deep Deterministic Policy Gradient (DDPG) and Twin Delayed Deep Deterministic Policy Gradient (TD3), under varying transmit power levels, RIS sizes, and learning hyperparameters. Simulation results demonstrate that SAC consistently achieves superior long-term sum-rate performance, particularly in low-to-moderate signal-to-noise ratio (SNR) regimes and for larger RIS dimensions, while exhibiting greater robustness to hyperparameter tuning. In contrast, DDPG and TD3 often converge faster in early training stages but are more sensitive to parameter settings and prone to suboptimal asymptotic solutions. These results establish SAC as a robust and scalable solution for RIS-assisted MU-MISO systems, offering improved reliability and adaptability for future intelligent wireless networks.
- Open Access
- 11 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 9696-9713
Twin Delayed Deep Deterministic Policy Gradient for Intelligent Optimization in STAR-RIS-Assisted Wireless Networks
Reconfigurable intelligent surfaces (RIS) have emerged as a key technology to enhance the performance of next-generation wireless networks by intelligently reconfiguring the propagation environment. In particular, simultaneously transmitting and reflecting RIS (STAR-RIS) extend this paradigm by enabling full-space coverage through concurrent reflection and transmission. This paper investigates a downlink multiple-input single-output (MISO) system assisted by a STAR-RIS and addresses the joint optimization of base station beamforming and RIS coefficients. To tackle the inherent non-convexity of this problem, we propose a reinforcement learning framework based on the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm. The proposed approach formulates the design task as a continuous control problem, allowing the agent to directly learn interference-aware policies that adapt to dynamic channel conditions. Extensive simulations validate the effectiveness of the proposed framework. The results demonstrate that the TD3-based policy achieves stable convergence and significantly improves the achievable sum rate compared to the baseline Deep Deterministic Policy Gradient (DDPG) method and the passive RIS benchmark. Performance scaling with the number of RIS elements and transmit power is clearly observed, confirming the scalability of the approach. In addition, hyperparameter sensitivity analysis highlights the importance of learning rate and decay parameter tuning for robust training. Cumulative distribution function (CDF) comparisons further show that the proposed framework enhances both average throughput and reliability across different channel realizations. The findings establish deep reinforcement learning, and TD3 in particular, as a promising tool for real-time optimization in STAR-RIS-assisted wireless systems. The proposed framework provides a flexible and scalable solution for intelligent resource allocation, paving the way for more reliable and efficient 6G communication networks.
- 11 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 9696-9713
Open AccessTwin Delayed Deep Deterministic Policy Gradient for Intelligent Optimization in STAR-RIS-Assisted Wireless Networks
Reconfigurable intelligent surfaces (RIS) have emerged as a key technology to enhance the performance of next-generation wireless networks by intelligently reconfiguring the propagation environment. In particular, simultaneously transmitting and reflecting RIS (STAR-RIS) extend this paradigm by enabling full-space coverage through concurrent reflection and transmission. This paper investigates a downlink multiple-input single-output (MISO) system assisted by a STAR-RIS and addresses the joint optimization of base station beamforming and RIS coefficients. To tackle the inherent non-convexity of this problem, we propose a reinforcement learning framework based on the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm. The proposed approach formulates the design task as a continuous control problem, allowing the agent to directly learn interference-aware policies that adapt to dynamic channel conditions. Extensive simulations validate the effectiveness of the proposed framework. The results demonstrate that the TD3-based policy achieves stable convergence and significantly improves the achievable sum rate compared to the baseline Deep Deterministic Policy Gradient (DDPG) method and the passive RIS benchmark. Performance scaling with the number of RIS elements and transmit power is clearly observed, confirming the scalability of the approach. In addition, hyperparameter sensitivity analysis highlights the importance of learning rate and decay parameter tuning for robust training. Cumulative distribution function (CDF) comparisons further show that the proposed framework enhances both average throughput and reliability across different channel realizations. The findings establish deep reinforcement learning, and TD3 in particular, as a promising tool for real-time optimization in STAR-RIS-assisted wireless systems. The proposed framework provides a flexible and scalable solution for intelligent resource allocation, paving the way for more reliable and efficient 6G communication networks.