Abstract: This paper presents a new method for graph isomorphism identification and applies it to the identification of isomorphic hybrid switching topology. A mathematical model of hybrid switching topology is first introduced. The adjacency matrix and undirected-weighted graph description of the hybrid switching
topology are presented according to the mathematical model. In this case, the problem of identifying isomorphic hybrid switching topology is transformed into isomorphism determination of the corresponding undirectedweighted graphs. An optimized version of the circuit simulation method previously presented by the authors is proposed to solve the graph isomorphism determination problem. With a small modification, the optimized circuit simulation method can solve the problem of graph isomorphism determination more efficiently. Tests of the proposed method and another method called the eigenvalue algorithm are applied in the identification of isomorphic hybrid switching topology. The results show that the optimized circuit simulation method is valid and has advantages in both identification efficiency and ability of matching corresponding vertices of the isomorphic hybrid switching topology.

Key words:  switching topology, isomorphism, circuit simulation method, eigenvalue algorithm