网格化量子漫步聚类模型及其算法分析  

Grid Changed Quantum Walk Clustering Model and its Algorithm Analysis

作　　者：黄德才[1] 钱国红[1] 陆亿红[1] 

机构地区：[1]浙江工业大学计算机科学与技术学院,杭州310023

出　　处：《小型微型计算机系统》2014年第4期802-806,共5页Journal of Chinese Computer Systems

基　　金：国家科技支撑计划子课题项目(2012BAD10B0101)资助;水利部公益性行业科研专项项目(201001031)资助

摘　　要：量子漫步作为一种新的量子计算模型具有巨大的前景,为经典算法寻求量子版本的解决方案提供了新思路.本文试图将量子漫步理论与聚类算法相结合,在分析图上离散量子漫步特点及其在解决聚类问题时存在不足的前提下,采取将漫步空间网格化的方式将模型简化,提出一种网格化量子漫步聚类模型,使之能够很好地完成聚类任务.该模型将数据点考虑为在量子网格世界中的根据特定规则执行漫步过程的量子,由于量子叠加等特性的存在,量子漫步聚类具有更好的时间效率和勘探能力,仿真实验也表明算法在聚类正确性上具有不错的表现.As a new quantum computing model, quantum walks has great prospects. It provides new idea for classical algorithms to find a solution of quantum version. This paper try to combine quantum walks and clustering. After analysis the characteristics of quantum walks on graph and its weakness in solving clustering problem, we simplify the model by making the space grid changed, and proposed a grid changed quantum walk clustering model which can solve the clustering problem perfectly. This model regard the data as quan- tum that is walking in the quantum grid world through special rules. Because of the quantum superposition principle, quantum walk clustering has a better efficience and exploration ability. Simulation results also show that algorithm on clustering correctness has a good performance.

关 键 词：网格 量子计算 量子漫步 数据聚类 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]