点簇聚合的目标顶点的优化  

Optimization of Representative Points for Vertex Clustering

作　　者：黄健民[1] 施法中[1] 

机构地区：[1]北京航空航天大学机械工程与自动化学院

出　　处：《系统仿真学报》2007年第20期4721-4724,共4页Journal of System Simulation

基　　金：广西自然科学基金(0447035)

摘　　要：研究了点簇聚合的目标顶点位置的计算问题。当计算过程中得到的目标顶点不在小单元之内,或者虽然在小单元之内,但目标顶点的位置不能唯一确定时,则将求解目标顶点的问题转化为求解带约束的二次优化问题。此二次优化问题的解既能保证目标顶点位于小单元之内,在位置上又最接近该点簇的重心。实验结果表明,该算法的时间效率类似于Lindstrom的算法,但在简化质量上要优于后者。The problem of calculating the representative points in vertex clustering was addressed. The problem of calculating the representative point was translated into the problem of constrained quadratic optimization when either the representative point obtained was outside of the cell or the representative point, which was though within the cell, and could not be uniquely determined. While securing the solution of the optimization problem wthin the cell, the position of the representative point is as close to the gravity of the cell as possible. Experimental results show that while the timing efficiency of the algorithm is similar to that of Lindstrom＇s, the simplification is much better than that of the latter＇s.

关 键 词：点簇聚合 目标顶点 位置优化 三维模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]