提出了多约束下连续体结构拓扑优化设计的K邻近(KNN)方法。首先以结构单元作为样本,将整个设计区域看作是样本空间,将连续体结构拓扑优化看作是一种对单元的模式识别;然后利用单元应力和位移灵敏度作为单元样本的特征向量,利用特征向量的欧氏距离为识别标准,引入模式识别的KNN方法进行多约束下的连续体结构拓扑优化设计;最后给出计算示例的优化结果来表明该方法是可行的。 A K-Nearest Neighbor (KNN) method for topological optimization of continuum structures with multiple constraints is proposed. Firstly, taking the structural unit as a sample, the entire design area is regarded as a sample space, and the continuum structure topology optimization is regarded as a pattern recognition of the unit. Then, the element stress and displacement sensitivity are used as the eigenvectors of the unit sample, The Euclidean distance of vector is used as identification standard, and the KNN method of pattern recognition is introduced to optimize the topology design of continuum structure with multiple constraints. Finally, the optimization results of calculation examples are given to show that this method is feasible.