超材料是人工设计的具有自然材料没有或很罕见的新奇特性。作为超材料的重要研究分支，具有特异力学功能的机械超材料研究成果层出不穷。然而，目前其主流设计依然是由灵感激发的尝试与修正，或通过大量数值优化来满足设计目标，不仅浪费时间，也需要巨大的计算资源。因此，这种设计策略不易实现自动化设计，难以为机械超材料研究提供普适性指导方案。本项目提出机械超材料模块设计概念，并寻求其符号回归算法实现方法，构筑一整套机械超材料的自动化设计流程。本项目一方面将符号回归技术应用于机械超材料的模块化设计，高效、自动地构筑具有指定功能的超材料，另一方面，以模块化为纽带，将超材料的基本构建元素作为一种新型的“符号”，把机械超材料作为回归对象，拓展了符号回归的运用空间，促进数学方法与工程应用的相互融合，具有重要的研究意义。

Metamaterials are artificially designed structural material with novel properties that are rare or even don’t exist in natural materials. As an important branch of metamaterials, mechanical metamaterials with special mechanical functionalities are continuously investigated. However, at present, its mainstream design is still inspired by trial-and-error method, or through a large number of numerical optimizations to satisfy design requirements, which are time wasting and computation consuming. Therefore, it is difficult for this strategy to achieve automated design, and it cannot provide a general guidance for mechanical metamaterials design. This project proposes the concept of modular design of mechanical metamaterials, and seeks to implement it by symbolic regression algorithm to construct a complete set of automated design for mechanical metamaterials. On the one hand, this project applies symbolic regression technology to the modular design of mechanical metamaterials to efficiently and automatically construct metamaterials with specified functions. On the other hand, using modular design as a link, and the basic construction elements of metamaterials as a new type of "symbol", mechanical metamaterials with different functions as regression objects, this project could expand the application space of symbol regression, and promote the integration of mathematical methods and engineering applications, which has important research significance.