提出一种基于虚节点的非度量加权多维标度定位算法,它利用矩阵截断奇异值分解计算节点相异性矩阵的逼近阵。仿真实验显示,该算法在网络节点密度较低或拓扑结构不规则时比以往算法有更好的定位精度。
This paper proposed a new nonmetric weighted MDS based on virtual node algorithm to solve the problem,it computed the approximate matrix of dissimilarity matrix by use of truncated singular value decomposition. Simulation results demonstrate that the new algorithm can promote localization accuracy in the low-density network node or irregular topology.