Computational techniques for vertex partitioning of graphs.
Computational techniques for vertex partitioning of graphs.
复制标题
图的顶点划分的计算技术。
DOI:
10.1021/ci00067a009
复制
发表时间:
1990
期刊:
影响因子:
--
通讯作者:
Munk,ME
中科院分区:
文献类型:
--
作者:
Liu,XY;Balasubramanian,K;Munk,ME
The graph-vertex-automorphism partitioning problem has received considerable attention in recent years. 1" 12 The vertex partitioning is of fundamental interest since it has many practical applications. First and foremost of all, it provides a solution for computer perception of the hidden topological symmetry of a molecule. In our group we have been interested in building a comprehensive computer-assisted structure-elucidation system. 11, 12 A critical problem encountered in this work is that given the neighborhood table (equivalently the adjacency matrix) of a molecule, can an automated algorithm and code be written to yield its topological symmetry. The graph-vertex-partitioning problem also finds important application in the computer generation of 13C and protonNMR signals of molecules and their intensity patterns. We have described in an earlier manuscript the application of the vertex partitioning to generate l3C NMR spectra. 24 Some early techniques for generatingvertex partitioning of graphs have been based on the Morgan algorithm. 1" 5 and the principal eigenvector algorithm. Randic and co-workers10 have formulated the canonical vertex-labeling methodto generate the vertex-automorphism partitionings of graphs. Although many of these algorithms and the codes based on these al-gorithms are simple to use, they often leadto convergence problems and oscillatory behaviors. Herndon and co-work-