Abstract
Reflectarray antennas (RA) offer a viable alternative to traditional antenna systems, merging the benefits of parabolic reflector antennas and phased array antennas. These antennas comprise a planar array of radiating elements stimulated by a feed antenna, allowing for precise radiation pattern manipulation. Despite their efficacy, their design involves intricate interactions between various variables, posing significant challenges. However, recent advancements in artificial intelligence (AI) and machine learning (ML) techniques are now being applied to circumvent these difficulties. Surrogate modelling, a method incorporating AI, enables the creation of computationally efficient approximations of complex systems like RA. Using ML techniques like regression, artificial neural networks, support vector machines, and Gaussian process regression, these surrogate models predict the RA's performance across different design parameters, thus enabling efficient exploration and optimisation of the design space. The integration of AI and ML methodologies can enhance RA design through computational efficiency, efficient design space exploration, and adaptive and data-driven design, potentially improving the performance and applicability of RAs in modern communication systems.