Recent Advances of Space-Mapping-Based Surrogate Models for Electromagnetic Optimization

This paper presents two representative space-mapping-based surrogate modeling strategies for accelerating electromagnetic (EM) optimization. The first approach constructs mesh-based coarse models through structural simplification and mesh deformation, while the second derives reduced-order models from fine simulations via projection-based model order reduction (MOR). Their modeling principles, optimization workflows, and performance in microwave filter design are systematically compared. The study highlights key trade-offs among accuracy, efficiency, and automation, offering practical guidelines for selecting suitable surrogate strategies in EM design tasks.