Composite planar coverings of graphs (Q1398265)

All graphs \(G= (V,E)\) considered in the present paper are simple and finite ones. A graph \(\widetilde G\) is called an (\(n\)-fold) covering of \(G\) with a projection \(p:\widetilde G\to G\) if there is an \(n\)-to-one surjection \(p: V(\widetilde G)\to V(G)\) which sends the neighbors of each \(v\in V(\widetilde G)\) bijectively to those in \(p(v)\). By an additional condition \(\widetilde G\) is defined to be a regular covering. In former papers the author investigated the relationship between planar 2-fold (or double) coverings and embeddings of graphs in the projective plane. These results (see Theorems 1 and 2) induced the following ``\(1-2-\infty\) conjecture'' or his ``planar cover conjecture'': A connected graph is projection-planar if and only if it has a planar covering. Many papers around this conjecture exist in the literature, but the sufficiency is still open; and therefore the author presents a new aspect of coverings of graphs. Namely, an \(n\)-fold covering or a covering projection \(p:\widetilde G\to G\) is said to be \((n_1,n_2)\)-composite if there are an \(n_1\)-fold covering \(p_1: \widetilde G\to G'\) and an \(n_2\)-fold covering \(p_2: G'\to G\) with \(p= p_1p_2\), and hence \(n= n_1n_2\). And if \(n_1> 1\) and \(n_2> 1\) then the covering is said to be composite. The main results of the paper are the following theorems: Theorem 4. A connected graph \(G\) is projective-planar if and only if it has an \((n,2)\)-composite planar connected covering for some \(n\geq 1\). Theorem 5. Every planar connected regular covering of a nonplanar connected graph is \((n,2)\)-composite for some \(n\geq 1\). In connection with the proofs of these results some structural properties of coverings are given, for example Theorem 12: Every connected regular covering of a 3-connected graph is either 3-connected or a cyclic chain.