网格空间映射电磁优化方法的加速策略研究进展OA

To address the challenges of complex coarse model construction,discontinuous responses,and low optimization efficiency in three-dimensional microwave device optimization,this paper reviews recent developments in multi-strategy accelerated mesh space mapping(MSM)techniques.Existing studies typically rely on coarse finite-element models and employ three key strategies to enhance modeling efficiency:firstly,sharp-feature approximation to simplify curved boundaries such as fillets and tuning screws and thereby reduce mesh density and geometric complexity;secondly,radial basis function-based mesh deformation to maintain response continuity and differentiability under geometric perturbations;finally,sensitivity-guided parameter screening to extract dominant correlated variables between coarse and fine models,thus reducing mapping dimensionality and training sample requirements.Building on these strategies,an integrated"sharpening-deformation-screening"framework for coarse-model construction has been established and applied in representative cases such as four-cavity tuned waveguide filters,demonstrating its potential for lowering optimization costs while maintaining accuracy.This review summarizes the applicability of these strategies in scenarios involving multi-resonant structures,high-dimensional parameter spaces,and strongly discrete responses,and provides methodological insights and future directions for developing automated and cost-effective electromagnetic optimization systems.