Artificial Intelligence for Antenna Design, Surrogate Modeling, And Optimization: A Review of Methods, Trends, And Open Challenges

Abstract: The increasing complexity of modern wireless communication systems has significantly transformed the antenna design process. Conventional antenna development primarily relies on full-wave electromagnetic (EM) simulations combined with iterative optimization techniques. Although these methods provide accurate results, they often require extensive computational time, particularly when dealing with multi-parameter, multi-objective, or broadband antenna structures. In recent years, artificial intelligence (AI) has emerged as a promising alternative for accelerating antenna analysis and design. Machine learning (ML), deep learning (DL), surrogate modeling, inverse design, and generative approaches have demonstrated the ability to predict antenna performance with considerably fewer electromagnetic simulations while maintaining satisfactory accuracy. This review presents a comprehensive analysis of recent research on AI-assisted antenna design, covering studies involving microstrip patch antennas, ultra-wideband (UWB) antennas, antenna arrays, metasurfaces, horn antennas, and reconfigurable antenna structures. The reviewed literature is organized according to major AI methodologies, including classical machine learning algorithms, deep learning architectures, surrogate modeling techniques, inverse design frameworks, and hybrid optimization approaches. Their advantages, limitations, and practical applications are critically discussed and compared. Furthermore, this review identifies several challenges that continue to limit the widespread adoption of AI-driven antenna design, including the absence of standardized benchmark datasets, inconsistent reporting of computational efficiency, limited experimental validation, and the difficulty of developing generalized models that can perform well across multiple antenna topologies. Finally, potential research directions such as physics-informed learning, transfer learning, generative artificial intelligence, digital twins, and automated antenna design frameworks are highlighted. The objective of this review is to provide researchers and engineers with a structured overview of current developments while identifying opportunities for future advancements in intelligent antenna design.

Keywords: Artificial Intelligence; Machine Learning; Deep Learning; Antenna Design; Surrogate Modeling; Inverse Design; Antenna Optimization; Electromagnetic Simulation; Metasurfaces; Wireless Communication