On regular graphs with \v{S}olt\'es vertices

From MaRDI portal
Publication:6430372



Abstract: Let W(G) be the Wiener index of a graph G. We say that a vertex vinV(G) is a v{S}olt'es vertex in G if W(G−v)=W(G), i.e. the Wiener index does not change if the vertex v is removed. In 1991, v{S}olt'es posed the problem of identifying all connected graphs G with the property that all vertices of G are v{S}olt'es vertices. The only such graph known to this day is C11. As the original problem appears to be too challenging, several relaxations were studied: one may look for graphs with at least k v{S}oltes vertices; or one may look for alpha-v{S}olt'es graphs, i.e. graphs where the ratio between the number of v{S}olt'es vertices and the order of the graph is at least alpha. Note that the original problem is, in fact, to find all 1-v{S}olt'es graphs. We intuitively believe that every 1-v{S}olt'es graph has to be regular and has to possess a high degree of symmetry. Therefore, we are interested in regular graphs that contain one or more v{S}olt'es vertices. In this paper, we present several partial results. We describe a construction that gives rise to an infinite family of cubic 2-connected graphs with two v{S}olt'es vertices. Moreover, we report that a computer search on publicly available collections of vertex-transitive graphs did not reveal any 1-v{S}olt'es graph. We are only able to provide examples of large frac13-v{S}olt'es graphs that are obtained by truncating certain cubic vertex-transitive graphs. This leads us to believe that no 1-v{S}olt'es graph other than C11 exists.














This page was built for publication: On regular graphs with \v{S}olt\'es vertices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6430372)