超材料的智能设计研究进展  

Advances in Intelligent Design of Metamaterials

作　　者：代铁琳 金刘超 尚宸 翟晓雅 傅孝明 刘利刚[1] Dai Tielin;Jin Liuchao;Shang Chen;Zhai Xiaoya;Fu Xiaoming;Liu Ligang(School of Mathematical Sciences,University of Science and Technology of China,Hefei 230026;Department of Mechanical and Automation Engineering,Chinese University of Hong Kong,Hong Kong 999077;School of Data Science,University of Science and Technology of China,Hefei 230026)

机构地区：[1]中国科学技术大学数学科学学院,合肥230026 [2]香港中文大学机械与自动化工程学系,中国香港999077 [3]中国科学技术大学大数据学院,合肥230026

出　　处：《计算机辅助设计与图形学学报》2025年第1期1-27,共27页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金青年基金(62402467);国家自然科学基金面上项目(62025207);中国科学院战略重点研究计划(XDB0640000);中央高校青年创新重点科研基金(YD0010002010);浙江大学计算机辅助设计与图形系统全国重点实验室开放课题(A2303).

摘　　要：超材料是一种人工合成材料,其结构经过精心设计和精密控制,表现出不同于自然材料的特殊性能,这些具有独特功能的超材料在光学、电磁学、声学、力学等领域具有重要的应用价值.然而,传统的超材料设计通常依赖于人工经验,导致设计周期长,缺乏灵活性和可调性,难以实现大规模快速设计与量产.高效设计与性能调控超材料已成为CAD、机械工程和材料科学的重要研究方向.近年来,随着人工智能算法的快速发展,智能算法在超材料设计领域中的应用越来越广泛,展现出高效优化、生成多样性结构、缩短设计周期等优势.文中对超材料的智能设计研究进展进行综述,首先介绍超材料的基本概念和发展历史;然后从实际需求出发,阐述了超材料的应用领域与设计问题;提出了智能算法的核心问题是数据的表示与数据集的构建,针对这2方面进行详细的阐述与对比分析;还介绍了智能优化算法的框架;最后总结超材料设计领域所面临的高质量数据集匮乏、多目标优化难题、高分辨率超材料的高效计算等挑战,并展望了该领域未来的发展趋势即面向多样化功能需求的“可表达”“可编辑”“可分析”“可优化”和“可制造”的研究.Metamaterials are artificially synthesized materials whose structures are meticulously designed and precisely controlled to exhibit unique properties not found in natural materials.These functional metamaterials have significant applications in fields such as optics,electromagnetics,acoustics,and mechanics.However,traditional metamaterial design often relies on human expertise,resulting in lengthy design cycles,a lack of flexibility,and limited tunability,which hinder the rapid design and mass production of metamaterials.Efficient design and performance tuning of metamaterials have become crucial research directions in mechanical engineering,and materials science.In recent years,with the rapid development of artificial intelligence algorithms,the application of intelligent algorithms in metamaterial design has been expanding,demonstrating advantages such as efficient optimization,generation of diverse structures,and shortened design cycles.This paper first introduces the basic concepts and development history of meta materials,and then discusses the application fields and design issues of metamaterials from the perspective of practical needs.The paper identifies the core issues of intelligent design algorithms as data representation and dataset construction,providing a detailed explanation and comparative analysis of these two aspects.Additionally,it reviews the framework of intelligent optimization algorithms.Finally,this paper highlights the challenges in metamaterial design,including the scarcity of high-quality datasets,the complexity of multi-objective optimization,and the computational demands of high-resolution metamaterial analysis.Additionally,it explores future trends in the field,emphasizing the development of metamaterials that are“expressible”“editable”“analyzable”“optimizable”and“manufacturable”to address diverse functional requirements.

关 键 词：超材料设计 智能算法 几何表示 结构优化 超材料数据集 

分 类 号：TP391.72[自动化与计算机技术—计算机应用技术]