Information Processing Letters 93 (2005) 99–102
www.elsevier.com/locate/ipl
Isometric-path numbers of block graphs
✩Jun-Jie Pan
a, Gerard J. Chang
b,c,∗aDepartment of Applied Mathematics, National Chiao Tung University, Hsinchu 30050, Taiwan bDepartment of Mathematics, National Taiwan University, Taipei 10617, Taiwan
cMathematics Division, National Center for Theoretical Sciences at Taipei, Old Mathematics Building,
National Taiwan University, Taipei 10617, Taiwan Received 4 November 2003
Communicated by A.A. Bertossi
Abstract
An isometric path between two vertices in a graph G is a shortest path joining them. The isometric-path number of G, denoted by ip(G), is the minimum number of isometric paths required to cover all vertices of G. In this paper, we determine exact values of isometric-path numbers of block graphs. We also give a linear-time algorithm for finding the corresponding paths.
2004 Elsevier B.V. All rights reserved.
Keywords: Isometric path; Block graph; Cut-vertex; Algorithms
1. Introduction
An isometric path between two vertices in a graph
G is a shortest path joining them. The isometric-path
number of G, denoted by ip(G), is the minimum
num-ber of isometric paths required to cover all vertices of G. This concept has a close relationship with the game of cops and robbers described as follows. The game is played by two players, the cop and the
rob-ber, on a graph. The two players move alternatively,
✩
Supported in part by the National Science Council under grant NSC90-2115-M002-024.
* Corresponding author.
E-mail address: gjchang@math.ntu.edu.tw (G.J. Chang).
starting with the cop. Each player’s first move con-sists of choosing a vertex at which to start. At each subsequent move, a player may choose either to stay at the same vertex or to move to an adjacent ver-tex. The object for the cop is to catch the robber, and for the robber is to prevent this from happening. Nowakowski and Winkler [7] and Quilliot [8] inde-pendently proved that the cop wins if and only if the graph can be reduced to a single vertex by succes-sively removing pitfalls, where a pitfall is a vertex whose closed neighborhood is a subset of the closed neighborhood of another vertex. As not all graphs are cop-win graphs, Aigner and Fromme [1] introduced the concept of the cop-number of a general graph G, denoted by c(G), which is the minimum number of
0020-0190/$ – see front matter 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ipl.2004.09.021
100 J.-J. Pan, G.J. Chang / Information Processing Letters 93 (2005) 99–102
cops needed to put into the graph in order to catch the robber. On the way to giving an upper bound for the cop-numbers of planar graphs, they showed that a single cop moving on an isometric path P guar-antee that after a finite number of moves the robber will be immediately caught if he moves onto P . Ob-serving this fact, Fitzpatrick [4] then introduced the concept of isometric-path cover and pointed out that
c(G) ip(G).
The isometric-path number of the Cartesian prod-uct Pn1× Pn2 × · · · × Pnd has been studied in the
lit-erature. Fitzpatrick [5] gave bounds for the case when
n1= n2= · · · = nd. Fisher and Fitzpatrick [3] gave
exact values for the case d= 2. Fitzpatrick et al. [6] gave a lower bound, which is in fact the exact value if d+ 1 is a power of 2, for the case when n1= n2=
· · · = nd= 2.
The purpose of this paper is to give exact values of isometric-path numbers of block graphs. We also give a linear-time algorithm to find the correspond-ing paths. For technical reasons, we consider a slightly more general problem as follows. Suppose every ver-tex v in the graph G is associated with a non-negative integer f (v). We call such function f a vertex
label-ing of G. An f -isometric-path cover of G is a family
C of isometric paths such that the following conditions
hold.
(C1) If f (v)= 0, then v is in an isometric path in C. (C2) If f (v) 1, then v is an end vertex of at least
f (v) isometric paths inC, while the counting is
twice if v itself is a path inC.
The f -isometric-path number of G, denoted by
ipf(G), is the minimum cardinality of an f
-isometric-path cover of G. It is clear that when f (v)= 0 for all vertices v in G, we have ip(G)= ipf(G). The attempt
of is paper is to determine the f -isometric-path num-ber of a block graph. Recall that a block graph is a graph in which every block is a complete graph. A
cut-vertex of a graph is a cut-vertex whose removal results in a
graph with more components than the original graph. It is well known that in a block graph all internal ver-tices of an isometric path are cut-verver-tices.
2. Block graphs
In this section, we determine the f -isometric-path numbers for block graphs G. Without loss of general-ity, we may assume that G is connected. First, a useful lemma.
Lemma 1. Suppose x is a non-cut-vertex of a block
graph G with a vertex labeling f . If vertex labeling
fis the same as f except that f(x)= max{1, f (x)},
then ipf(G)= ipf(G).
Proof. As any internal vertex of an isometric path in
a block graph is a cut-vertex but x not a cut-vertex,
x must be an end vertex of any isometric path. It
fol-lows that a collectionC is an f -isometric-path cover if and only if it is an f-isometric-path cover. The lemma then follows. 2
So, now we may assume that f (v) 1 for all non-cut-vertices v of G, and call such a vertex labeling
regular. Now, we have the following theorem for the
inductive step.
Theorem 2. Suppose G is a block graph with a regular
labeling f , and x is a non-cut-vertex in a block B with exactly one cut-vertex y or with no cut-vertex in which case let y be any vertex of B− {x}. When f (x) = 1, let
G= G − x with a regular vertex labeling fwhich is the same as f except f(y)= f (y) + 1. When f (x)
2, let G= G with a regular vertex labeling fwhich is the same as f except f(x)= f (x) − 1 and f(y)=
f (y)+ 1. Then ipf(G)= ipf(G).
Proof. We first prove that ipf(G) ipf(G).
Sup-pose C is an optimal f -isometric-path cover of G.
Choose a path P inC having x as an end vertex. We consider four cases.
Case 1.1. P = x and f (x) = 1 (i.e., G= G − x).
In this case, C= (C − {P }) ∪ {y} is an f -isomet-ric-path cover of G. Hence, ipf(G)= |C| |C|
ipf(G).
Case 1.2. P = x and f (x) 2 (i.e., G= G). In
this case, C= (C − {P }) ∪ {xy} is an f -isometric-path cover of G. Hence, ipf(G) = |C| |C|
J.-J. Pan, G.J. Chang / Information Processing Letters 93 (2005) 99–102 101
Case 1.3. P= xz for some vertex z in B −{x, y}. In
this case,C= (C −{P })∪{xy} is an f-isometric-path cover of G. Hence, ipf(G)= |C| |C| ipf(G).
Case 1.4. P= xyQ, where Q contains no vertices
in B. In this case,C= (C − {P }) ∪ {yQ} is an f -iso-metric-path cover of G. Hence, ipf(G)= |C| |C|
ipf(G).
Next, we prove that ipf(G) ipf(G). SupposeC
is an optimal f-isometric-path cover of G. Choose a path PinC having y as an end vertex. We consider three cases.
Case 2.1. P= yx. In this case, G= G and C =
(C− {P}) ∪ {x} is an f -isometric-path cover of G.
Hence, ipf(G) |C| |C| = ipf(G).
Case 2.2. P= yz for some z in B − {x, y}. In this
case, C = (C− {P}) ∪ {xz} is an f -isometric-path cover of G. Hence, ipf(G) |C| |C| = ipf(G).
Case 2.3. P= yQ, where Q contains no vertex in
B. In this case,C = (C− {P}) ∪ {xyQ} is an f
-iso-metric-path cover of G. Hence, ipf(G) |C| |C| =
ipf(G).
Consequently, we have the following result for
f -isometric-path numbers of connected block graphs.
Theorem 3. If G is a connected block graph with a
regular vertex labeling f , then ipf(G)= s(G)/2,
where s(G)=v∈V (G)f (v).
Proof. The theorem is obvious when G has only one
vertex. For the case when G has more than one ver-tex, we apply Theorem 2 repeatedly until the graph becomes trivial. Notice that s(G)= s(G) when
The-orem 2 is applied. 2
For the isometric-path-cover problem, we have
Corollary 4. If G is a connected block graph, then
ip(G)= nc(G)/2, where nc(G) is the number of
non-cut-vertices of G.
Proof. The corollary follows from Theorem 3 and the
fact that ip(G)= ipf(G) for the regular vertex
la-beling f with f (v)= 1 if v is a non-cut-vertex and
f (n)= 0 otherwise. 2
3. Algorithm
Based on Theorem 2, we are able to design an al-gorithm for the isometric-path-cover problem in block graphs. Notice that we may only consider connected block graphs with regular vertex labelings. To speed up the algorithm, we may modify Theorem 2 a little bit so that each time a non-cut-vertex is handled.
Theorem 5. Suppose G is a block graph with a regular
labeling f , and x is a non-cut-vertex in a block B with exactly one cut-vertex y or with no cut-vertex in which let y be any vertex in B− {x}. Let G= G − x with a regular vertex labeling f which is the same as f except f(y)= f (y)+f (x). Then ipf(G)= ipf(G).
Proof. The theorem follows from repeatedly applying
Theorem 2. 2
Now, we are ready to give the algorithm.
Algorithm PG. Find the f -isometric-path number ipf(G) of a connected block graph.
Input. A connected block graph G and a regular vertex la-beling f .
Output. An optimal f -isometric-path cover C of G and ipf(G).
Method.
1. construct a stack S which is empty at the beginning; 2. let G← G;
3. while (Ghas more than one vertex) do
4. choose a block B with exactly one cut-vertex y or with no cut-vertex in which case choose any y∈ B; 5. for (all vertices x in B− {y}) do
6. f (y)← f (y) + f (x); 7. push (x, y, f (x)) into S; 8. G← G− x;
9. end for; 10. end while;
11. ipf(G)← f (r)/2, where r is the only vertex of G; 12. letC be the family of isometric paths containing ip(G)
copies of the path r ; 13. while (S is not empty) do 14. pop (x, y, i) from S;
15. choose i copies of path P inC using y as an end vertex;
16. if (P= yx) then
102 J.-J. Pan, G.J. Chang / Information Processing Letters 93 (2005) 99–102
18. if (P= yz for some vertex z in the block of G con-taining x) then
19. replace the i copies of P by i copies of xz inC; 20. if (P= yQ where Q has no vertices in the block of
G containing x) then
21. replace the i copies of P by the i copies of xyQ inC;
22. end while.
Algorithm PG can be implemented in time linear to the number of vertices and edges. Notice that we can use the depth-first search to find all blocks and cut-vertices of a graph in linear time, see [2].
Acknowledgements
The authors thank the referee for many constructive suggestions.
References
[1] M. Aigner, M. Fromme, A game of cops and robbers, Discrete Appl. Math. 8 (1984) 1–12.
[2] T.H. Cormen, C.E. Leiserson, R.L. Rivest, Introduction to Algo-rithms, MIT Press, Cambridge, MA, 1990.
[3] D.C. Fisher, S.L. Fitzpatrick, The isometric path number of a graph, J. Combin. Math. Combin. Comput. 38 (2001) 97–110. [4] S.L. Fitzpatrick, Aspects of domination and dynamic
domina-tion, Ph.D. Thesis, Dalhousie University, Nova Scotia, Canada, 1997.
[5] S.L. Fitzpatrick, The isometric path number of the Cartesian product of paths, Congr. Numer. 137 (1999) 109–119. [6] S.L. Fitzpatrick, R.J. Nowakowski, D. Holton, I. Caines,
Cov-ering hypercubes by isometric paths, Discrete Math. 240 (2001) 253–260.
[7] R. Nowakowski, P. Winkler, Vertex-to-vertex pursuit in a graph, Discrete Math. 43 (1983) 235–239.