集群环境下的并行聚类算法

探讨在集群环境下，如何设计并行聚类算法。作为一种低成本、通用并行系统，集群系统的通信能力，相对于节点的计算能力，是一个瓶颈。所以在集群环境下，设计并行聚类算法时，应采用数据并行的思想。从理论上，对采用数据并行思想后，影响聚类算法的加速比和聚类质量的因素进行了分析，然后通过一个验证算法PCIT(Parallel clustering algorithm based on Index Tree)证实了理论分析的正确性。研究结果可以为以后设计更好的数据并行聚类算法提供理论依据。

Parallel Clustering Algorithm for PCs Cluster

The goal of this paper is to discuss how to design a parallel clustering algorithm for PCs cluster. As a kind of low-cost all-purpose parallel computing system, the bottleneck of PCs cluster is the communication ability compared to the computing ability. This paper proposes the idea of adopting data parallelism when designing a parallel clustering algorithm for PCs cluster. The paper theoretically analyzes the influence on the speedup and the quality of clustering firstly. And then it uses the experimental result of an algorithm named PCIT(parallel clustering algorithm based on index tree), to prove the correctness of the theoretical analysis. This theoretical analysis will be helpful to design better parallel clustering algorithm for PCs cluster.