Data-driven non-intrusive shape-topology collaborative optimization method and system

The invention discloses a data-driven non-intrusive shape-topology collaborative optimization method and system, and relates to the field of structure optimization. According to the method, firstly, parametric modeling is conducted on a structure through the grid deformation technology, displacement of control points of the structure is restrained through a shape equation, and it is guaranteed that the structure shape is regular and smooth after grid deformation; secondly, carrying out topological optimization on the models with different shapes by adopting various solvers to obtain multi-source topological optimization response data; and finally, establishing a non-intrusive agent model between the shape design variable and the topological optimization response data, and obtaining a shape-topological collaborative optimization scheme based on the non-intrusive agent model. According to the shape-topology optimization method, the interaction between shape optimization and topology optimization is considered,