摘要
为了减小最短路径距离矩阵与欧氏距离矩阵之间的差异,提高MDS-MAP(C)算法的节点定位精度,提出一种改进的多维标度节点定位算法。该算法对MDS-MAP(C)算法进行了以下改进:采用启发式的搜索策略对最短路径距离矩阵进行修正,以减少最短路径距离矩阵与实际的欧氏距离矩阵之间的误差;利用smacof算法迭代误差函数代替SVD分解来求解节点的定位问题,以优化和改善节点定位的求解过程。实验结果表明,与MDS-MAP(C)算法相比,改进算法能够减少最短路径距离的误差,有效提高节点的定位精度,并且对不规则网络具有更好的适应性。
In order to reduce the difference between the shortest path distance matrix and Euclidean distance ma-trix,an improved algorithm of multidimensional scaling node localization was proposed to enhance the node localiza-tion accuracy of MDS-MAP（C）algorithm. The algorithm made some improvements on MDS-MAP（C）algorithm. Theshortest path distance matrix was corrected by using heuristic search strategy,so as to reduce the error between theshortest path distance matrix and the actual Euclidean distance matrix. Then smacof algorithm iterative error func-tion instead of singular value decomposition（SVD）was utilized to solve the problem of node localization,whichcould optimize and improve the solving process of node localization. The experimental results show that comparedwith MDS-MAP（C）algorithm,the improved algorithm can reduce the error of the shortest path distance,effectivelyimprove the node localization accuracy,and it has better adaptability to the irregular network.
出处
《传感技术学报》
CAS
CSCD
北大核心
2016年第1期129-135,共7页
Chinese Journal of Sensors and Actuators
