Process model clustering is a hot topic in the field of process management. This paper presents a new K-means algorithm based on cuckoo algorithm, which compensates drawbacks of traditional K-means algorithm, such as relying on initial solution and being easily trapped in local optimums. In this paper, simulated data sets consist of five features (process model structure performance, cost, efficiency, customer satisfaction and quality), but experiments are conducted by only two indicators with higher weight. Experimental results show that the method has relatively higher feasibility and effectiveness.%流程模型聚类是流程管理领域的一个热门话题。本文提出一种基于布谷鸟算法的K-means算法，该算法弥补了K-means算法的依赖初始解、易陷入局部最优等缺点。本文从流程模型结构性能、成本、效率、顾客满意度以及质量等五个方面模拟数据集，并选择权重较高的属性进行试验操作，结果表明算法的具有较高的可行性和有效性。
展开▼
