The construction of graph G k in the proof of Proposition 5, with an optimal LTD-set (black vertices). Let C be an LTD-set of G k . Consider the k − 1 closed neighborhoods N [s i ] for 1 ≤ i < k. If we have |N [s i ] ∩ C| ≥ 2 for all i with 1 ≤ i < k, then 1≤i<k (N [s i ] ∩ C) ≥ 2k − 2, and at least one of the remaining vertices s k , q k , q ′ k must belong to C, as otherwise N (q k ) ∩ C = N (q ′ k ) ∩ C would follow, a contradiction. This implies |C| ≥ 2k − 1. If, however, for some i with 1 ≤ i < k, we have |N [s i ] ∩ C| = 1, then s i / ∈ C since otherwise s i is not totally dominated by the set C. If N [s i ] ∩ C = {q i }, then N (q ′ i ) ∩ C = Q ∩ C. If N [s i ] ∩ C = {q ′ i }, then N (q i ) ∩ C = (Q ∪ {s k }) ∩ C. The two possibilities can occur at most once each. Assume that they both occur once each, with N [s a ] ∩ C = {q ′ a } and N [s b ] ∩ C = {q b } (with 1 ≤ a < b < k). Note that s k ∈ C, otherwise q a and q ′ b are not located. Moreover, C must contain q ′ k (otherwise q ′ b and q ′ k are not located) and q k (otherwise q a and q k are not located), and so |C| ≥ 2k − 1, as claimed. Similarly, if we have |N [s i ] ∩ C| ≥ 2 for all i with 1 ≤ i < k except that N [s a ] ∩ C = {q ′ a }, if s k ∈ C, then q k ∈ C, otherwise q a and q k are not located. If s k / ∈ C, then both q k , q ′ k are in C to locate the vertices q a , q k , q ′ k . Thus, again |C| ≥ 2k − 1. Finally, if we have |N [s i ] ∩ C| ≥ 2 for all i with 1 ≤ i < k except that N [s b ] ∩ C = {q b }, if s k ∈ C, then q ′ k ∈ C, otherwise q ′ b and q ′ k are not located. If s k / ∈ C, then both q k , q ′ k are in C to locate the vertices q ′ b , q k , q ′ k , and again |C| ≥ 2k − 1. Thus, in all the above cases, we have |C| ≥ 2k − 1 and, together with the upper bound γ L t (G k ) <

The construction of graph G k in the proof of Proposition 5, with an optimal LTD-set (black vertices). Let C be an LTD-set of G k . Consider the k − 1 closed neighborhoods N [s i ] for 1 ≤ i < k. If we have |N [s i ] ∩ C| ≥ 2 for all i with 1 ≤ i < k, then 1≤i<k (N [s i ] ∩ C) ≥ 2k − 2, and at least one of the remaining vertices s k , q k , q ′ k must belong to C, as otherwise N (q k ) ∩ C = N (q ′ k ) ∩ C would follow, a contradiction. This implies |C| ≥ 2k − 1. If, however, for some i with 1 ≤ i < k, we have |N [s i ] ∩ C| = 1, then s i / ∈ C since otherwise s i is not totally dominated by the set C. If N [s i ] ∩ C = {q i }, then N (q ′ i ) ∩ C = Q ∩ C. If N [s i ] ∩ C = {q ′ i }, then N (q i ) ∩ C = (Q ∪ {s k }) ∩ C. The two possibilities can occur at most once each. Assume that they both occur once each, with N [s a ] ∩ C = {q ′ a } and N [s b ] ∩ C = {q b } (with 1 ≤ a < b < k). Note that s k ∈ C, otherwise q a and q ′ b are not located. Moreover, C must contain q ′ k (otherwise q ′ b and q ′ k are not located) and q k (otherwise q a and q k are not located), and so |C| ≥ 2k − 1, as claimed. Similarly, if we have |N [s i ] ∩ C| ≥ 2 for all i with 1 ≤ i < k except that N [s a ] ∩ C = {q ′ a }, if s k ∈ C, then q k ∈ C, otherwise q a and q k are not located. If s k / ∈ C, then both q k , q ′ k are in C to locate the vertices q a , q k , q ′ k . Thus, again |C| ≥ 2k − 1. Finally, if we have |N [s i ] ∩ C| ≥ 2 for all i with 1 ≤ i < k except that N [s b ] ∩ C = {q b }, if s k ∈ C, then q ′ k ∈ C, otherwise q ′ b and q ′ k are not located. If s k / ∈ C, then both q k , q ′ k are in C to locate the vertices q ′ b , q k , q ′ k , and again |C| ≥ 2k − 1. Thus, in all the above cases, we have |C| ≥ 2k − 1 and, together with the upper bound γ L t (G k ) <

Source publication
Preprint
Full-text available
We study upper bounds on the size of optimum locating-total dominating sets in graphs. A set $S$ of vertices of a graph $G$ is a locating-total dominating set if every vertex of $G$ has a neighbor in $S$, and if any two vertices outside $S$ have distinct neighborhoods within $S$. The smallest size of such a set is denoted by $\gamma^L_t(G)$. It has...

Contexts in source publication

Context 1
... that q ′ k has no neighbor in S and that the sets N (s i ) are disjoint for 1 ≤ i < k. See Figure 1 for an illustration. ...
Context 2
... n ≥ 7. In what follows, we adopt the notation that if there is a (d 1 , d 2 , . . . , d k )-sequence in G, then P : v 1 v 2 . . . ...
Context 3
... of subclaim. (a) Suppose that there is a (1, 3, 1)-sequence in G. In this case, n ′ = n − 3 ≥ 4. ...
Context 4
... that there is a (1, 2, 2)-sequence in G. Let v ′ be the second neighbor of v 3 . As in the previous case, n ′ = n − 3 ≥ 4 and G ′ is connected. ...
Context 5
... in the previous case, n ′ = n − 3 ≥ 4 and G ′ is connected. By part (a), there is no (1, 3, 1)-sequence, implying that G ′ / ∈ F tdom and γ L t (G ′ ) ≤ 2 3 n ′ = 2 3 n − 2. As before every γ L t -set of G ′ can be extended to a LTD-set of G by adding to it the vertices v 2 and v 3 , implying that γ L t (G) ≤ 2 3 n, a contradiction. (c) Suppose that there is a (1, 2, 3, 1)-sequence in G. ...
Context 6
... part (a), there is no (1, 3, 1)-sequence, implying that G ′ / ∈ F tdom and γ L t (G ′ ) ≤ 2 3 n ′ = 2 3 n − 2. As before every γ L t -set of G ′ can be extended to a LTD-set of G by adding to it the vertices v 2 and v 3 , implying that γ L t (G) ≤ 2 3 n, a contradiction. (c) Suppose that there is a (1, 2, 3, 1)-sequence in G. In this case, n ′ = n − 4 ≥ 3 and G ′ is connected. ...
Context 7
... this case, n ′ = n − 4 ≥ 3 and G ′ is connected. By part (a), there is no (1, 3, 1)-sequence, implying that G ′ / ∈ F tdom and γ L t (G ′ ) ≤ 2 3 n ′ = 2 3 (n−4) < 2 3 n−2. Every γ L t -set of G ′ can be extended to a LTD-set of G by adding to it the vertices v 2 and v 3 , implying that γ L t (G) ≤ 2 3 n, a contradiction. ...
Context 8
... γ L t -set of G ′ can be extended to a LTD-set of G by adding to it the vertices v 2 and v 3 , implying that γ L t (G) ≤ 2 3 n, a contradiction. (d) Suppose that there is a (1, 2, 3, 2, 1)-sequence in G. In this case, G ′ is connected and n ′ = n−5 ≥ 2. If G ′ ∈ F tdom , then G ′ ∼ = K 2 by the fact that there is no (1, 3, 1)-sequence in G by part (a). ...
Context 9
... Suppose that there is a (1, 2, 3, 2, 1)-sequence in G. In this case, G ′ is connected and n ′ = n−5 ≥ 2. If G ′ ∈ F tdom , then G ′ ∼ = K 2 by the fact that there is no (1, 3, 1)-sequence in G by part (a). The graph G is therefore determined, and is obtained from a star K 1,3 by subdividing every edge once. ...
Context 10
... Suppose that there is a (1, 2, 3)-sequence in G. In this case, n ′ = n− 3 ≥ 4 and G ′ contains at most two components. ...
Context 11
... γ L t -set of G ′ can be extended to a LTD-set of G by adding to it the vertices v 2 and v 3 , implying that γ L t (G) ≤ 2 3 n, a contradiction. (f) Since there is no (1, 2, 1)-sequence (since n ≥ 7), no (1, 2, 2)-sequence by (b) and no (1, 2, 3)-sequence by (e), there can be no (1, 2)-sequence in G. Hence, part (f) follows immediately from parts (b) and (e). ...
Context 12
... Suppose that there is a (1, 3, 2)-sequence in G. In this case, n ′ = n − 3 ≥ 4. If G ′ is disconnected, then by parts (a)-(f), neither component of G ′ belongs to F tdom . ...
Context 13
... Suppose that there is a (1, 3, 3, 1)-sequence in G. In this case, n ′ = n − 4 ≥ 3 and G ′ contains at most two components. ...
Context 14
... Subclaim H.1(f), the neighbor of every vertex of degree 1 has degree 3 in G. Further by Subclaim H.1(g), such a vertex of degree 3 has both its other two neighbors of degree 3. Therefore the existence of a vertex of degree 1 implies that there is a (1, 3, 3)-sequence in G. In this case, n ′ = n − 3 ≥ 4. Let u 2 be the neighbor of v 2 not on P , and let u 3 and w 3 be the two neighbors of v 3 not on P . ...

Similar publications

Preprint
Full-text available
A book embedding of a complete graph is a spatial embedding whose planar projection has the vertices located along a circle, consecutive vertices are connected by arcs of the circle, and the projections of the remaining "interior" edges in the graph are straight line segments between the points on the circle representing the appropriate vertices. A...
Preprint
Full-text available
Given two $k$-dicolourings of a digraph $D$, we prove that it is PSPACE-complete to decide whether we can transform one into the other by recolouring one vertex at each step while maintaining a dicolouring at any step even for $k=2$ and for digraphs with maximum degree $5$ or oriented planar graphs with maximum degree $6$. A digraph is said to be $...
Preprint
Full-text available
The fractional differential equation $L^\beta u = f$ posed on a compact metric graph is considered, where $\beta>\frac14$ and $L = \kappa - \frac{\mathrm{d}}{\mathrm{d} x}(H\frac{\mathrm{d}}{\mathrm{d} x})$ is a second-order elliptic operator equipped with certain vertex conditions and sufficiently smooth and positive coefficients $\kappa,H$. We de...
Preprint
Full-text available
An edge-coloring of a graph $G$ with colors $1,\ldots,t$ is called an \emph{interval $t$-coloring} if all colors are used and the colors of edges incident to each vertex of $G$ are distinct and form an interval of integers. In 1990, Kamalian proved that if a graph $G$ with at least one edge has an interval $t$-coloring, then $t\leq 2|V(G)|-3$. In 2...
Article
Full-text available
Let t, k, d be integers with 1≤d≤k-1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 \le d \le k - 1$$\end{document} and t≥8d+11\documentclass[12pt]{minimal} \usepacka...