The Decycling Number on Graphs and Digraphs
Min-Yun Lien
Advisor: Hung-Lin Fu
Department of Applied Mathematics, National Chiao Tung University, Taiwan
2014-08-03
Outline
1
Decycling Problem and Applications
2
Decycling Number and Cycle Packing Number
3
Decycling Number on Graphs
Path Product
4
Decycling Number on Digraphs
Generalized Kautz Digraph Generalized de Bruijn Digraphs
5
Objective Work
Min-Yun Lien 2/56
Definition and Applications
Definition (Decycling Problem)
Given a directed/undirected graphG= (V, E), find a minimum set D⊂ V such thatG\ Dis acyclic.
◮
H as applications in
Deadlock prevention in operating systems (Wang et al.
1985; Silberschatz et al. 2003)
Definition and Applications
Definition (Decycling Problem)
Given a directed/undirected graphG= (V, E), find a minimum set D⊂ V such thatG\ Dis acyclic.
◮
H as applications in
Deadlock prevention in operating systems (Wang et al.
1985; Silberschatz et al. 2003)
Example: An operating system schedules different processes.
A B
C
Process A is waiting for the resource on Process B so it can’t release its own resource.
Min-Yun Lien 3/56
Definition and Applications
Definition (Decycling Problem)
Given a directed/undirected graphG= (V, E), find a minimum set D⊂ V such thatG\ Dis acyclic.
◮
H as applications in
Deadlock prevention in operating systems (Wang et al.
1985; Silberschatz et al. 2003)
Example: An operating system schedules different processes.
B A
C Process A is waiting for no resource so it can release its resource.
Definition and Applications
Definition (Decycling Problem)
Given a directed/undirected graphG= (V, E), find a minimum set D⊂ V such thatG\ Dis acyclic.
◮
H as applications in
Deadlock prevention in operating systems (Wang et al.
1985; Silberschatz et al. 2003)
Example: An operating system schedules different processes.
B A
C
Deadlock:
Competing actions are each waiting for the other to finish, and thus neither ever does.
Solution:
Remove some processes to break such cycles and put them in a waiting queue.
Min-Yun Lien 5/56
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
Min-Yun Lien 6/56
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
⊲ Problem: Set the minimum number of vertices YES at beginning such that all vertices become YES eventually.
Min-Yun Lien 6/56
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
⊲ Problem: Set the minimum number of vertices YES at beginning such that all vertices become YES eventually.
⊲ A decycling set could be the only choice!
Example: A4-regular graph (such as toroidal mesh network).
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
⊲ Problem: Set the minimum number of vertices YES at beginning such that all vertices become YES eventually.
⊲ A decycling set could be the only choice!
Example: A4-regular graph (such as toroidal mesh network).
Decycling Set Decycling Set
Min-Yun Lien 7/56
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
⊲ Problem: Set the minimum number of vertices YES at beginning such that all vertices become YES eventually.
⊲ A decycling set could be the only choice!
Example: A4-regular graph (such as toroidal mesh network).
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
⊲ Problem: Set the minimum number of vertices YES at beginning such that all vertices become YES eventually.
⊲ A decycling set could be the only choice!
Example: A4-regular graph (such as toroidal mesh network).
Decycling Set Decycling Set
Min-Yun Lien 9/56
Applications
Monopolies in synchronous distributed systems (Peleg 1998; Peleg 2002)
⊲ Vertices are coloredYESorNO.
⊲ Monotonesynchronous system: at each step aNOvertex changes toYESif more than half of its neighbors areYES.
⊲ Problem: Set the minimum number of vertices YES at beginning such that all vertices become YES eventually.
⊲ A decycling set could be the only choice!
Example: A4-regular graph (such as toroidal mesh network).
NOvertices on the cycle remain NO.
Failed!
Applications
Constraint satisfaction problem (Dechter 1990).
Bayesian inference in artificial intelligence (Bar-Yehuda et al. 1998).
Converters’ placement problem in optical networks (Kleinberg and Kumar 1999).
VLSI chip design (Festa et al. 2000).
Min-Yun Lien 11/56
Complexity
◮
H as been extensively studied.
NP-hard
Reduction from VERTEX COVER (R. Karp 1972).
Even for planar graphs, bipartite graphs.
Related Problems
◮
I s related or equivalent to
Feedback vertex set problem(Wang et al. 1985).
Hitting cycle problem
Min-Yun Lien 13/56
Related Problems
◮
I s related or equivalent to
Feedback vertex set problem(Wang et al. 1985).
Hitting cycle problem
Maximum induced forest problem
(Erd ¨os, Saks and S ´os 1986 Maximum induced trees in graphs).
Variations
◮
H as the following variations (survey Festa et al. 2000).
Graph Bipartization Problem
FindD⊂ Esuch thatG\ Dhas no odd cycle.
Min-Yun Lien 14/56
Variations
◮
H as the following variations (survey Festa et al. 2000).
Graph Bipartization Problem
FindD⊂ Esuch thatG\ Dhas no odd cycle.
Weighted Decycling Problem
Seek a minimum-weight decycling setDwhere each vertex has a weight.
Variations
◮
H as the following variations (survey Festa et al. 2000).
Graph Bipartization Problem
FindD⊂ Esuch thatG\ Dhas no odd cycle.
Weighted Decycling Problem
Seek a minimum-weight decycling setDwhere each vertex has a weight.
Loop Cut Set Problem
GivenA(C) ⊆ V(C)for each cycleCofG, find a minimum setDsuch thatD∩ A(C) 6= ∅.
Min-Yun Lien 14/56
Relations with Cycle Packing Number
◮
I s often compared with the following graph parameter.
Cycle Packing Numberν(G): the maximum number of vertex-disjoint cycles ofG.
1 ) ( G Q
Definition:∇(G) :=the decycling number (minimum size of decycling set) ofG.
Relations with Cycle Packing Number
⊲ Dirac and Gallai had interest in the relations betweenν(G) and∇(G).
⊲ It is clear thatν(G) ≤ ∇(G).
Min-Yun Lien 16/56
Relations with Cycle Packing Number
⊲ Dirac and Gallai had interest in the relations betweenν(G) and∇(G).
⊲ It is clear thatν(G) ≤ ∇(G).
Definition:∇(k) := max{∇(G) : ν(G) = k}.
Relations with Cycle Packing Number
⊲ Dirac and Gallai had interest in the relations betweenν(G) and∇(G).
⊲ It is clear thatν(G) ≤ ∇(G).
Definition:∇(k) := max{∇(G) : ν(G) = k}.
⊲ ∇(1) ≤ 3;ν(K5) = 1and∇(K5) = 3(Bollob´as 1964).
⊲ ∇(2) = 6and9≤ ∇(3) ≤ 12(Voss 1968).
Min-Yun Lien 16/56
Relations with Cycle Packing Number
⊲ Dirac and Gallai had interest in the relations betweenν(G) and∇(G).
⊲ It is clear thatν(G) ≤ ∇(G).
Definition:∇(k) := max{∇(G) : ν(G) = k}.
⊲ ∇(1) ≤ 3;ν(K5) = 1and∇(K5) = 3(Bollob´as 1964).
⊲ ∇(2) = 6and9≤ ∇(3) ≤ 12(Voss 1968).
⊲ c1k log k≤ ∇(k) ≤ c2k log kfor some constantsc1 andc2
(Erd¨os and P´osa 1964).
Relations with Cycle Packing Number
Consider planar graphs:
Min-Yun Lien 17/56
Relations with Cycle Packing Number
Consider planar graphs:
Jones’ Conjecture (Kloks, Lee and Liu 2002) For every planar graphG,∇(G) ≤ 2ν(G).
Relations with Cycle Packing Number
Consider planar graphs:
Jones’ Conjecture (Kloks, Lee and Liu 2002) For every planar graphG,∇(G) ≤ 2ν(G).
Theorem (Chen, Fu and Shih 2010, TJM) For every planar graphG,∇(G) ≤ 3ν(G).
Min-Yun Lien 17/56
Decycling number of outerplanar graphs
Consider outerplanar graphs:
Theorem (Kloks, Lee and Liu 2002)
For every outerplanar graphG,∇(G) ≤ 2ν(G).
◮ An outerplanar graphGis calledlower-extremalif∇(G) = ν(G) andupper-extremalif∇(G) = 2ν(G).
Decycling number of outerplanar graphs
Consider outerplanar graphs:
Theorem (Kloks, Lee and Liu 2002)
For every outerplanar graphG,∇(G) ≤ 2ν(G).
◮ An outerplanar graphGis calledlower-extremalif∇(G) = ν(G) andupper-extremalif∇(G) = 2ν(G).
◮ Upper-Extremal Results:
We define asun graphS3as follows.
∇(S3) = 2 = 2ν(S3). S
Min-Yun Lien 18/56
Decycling number of outerplanar graphs
Theorem (Chang, Fu, Lien, 2011, JCO)
An outerplanar graphGis upper-extremal if and only ifGis anS3-tree.
A graph is anS3-tree of ordertif it has exactlyt
vertex-disjointS3-subdivisions and every edge not on these S3-subdivisions belongs to no cycle.
Decycling number of outerplanar graphs
Theorem (Chang, Fu, Lien, 2011, JCO)
An outerplanar graphGis upper-extremal if and only ifGis anS3-tree.
A graph is anS3-tree of ordertif it has exactlyt
vertex-disjointS3-subdivisions and every edge not on these S3-subdivisions belongs to no cycle.
Example:
An S3-tree G of order 3, where(G) 6 2Q(G).
Min-Yun Lien 19/56
Decycling number of outerplanar graphs
◮ Lower-Extremal Results:
The following graphs are NOT lower-extremal (∇(G) 6= ν(G)):
Sun graphsSkwith odd numberk:
S3 S5
∇(Sk) = ⌈k2⌉andν(Sk) = ⌊k2⌋.
Decycling number of outerplanar graphs
Theorem (Chang, Fu, Lien, 2011, JCO)
For an outerplanar graphG, ifGhas noSk-subdivision for all odd numberk, thenGis lower-extremal.
Min-Yun Lien 21/56
Decycling number of Graphs
Lower Bound of Undirected Graphs
Lemma (Beineke, 1997, JGT)
IfGis a connected graph withpvertices (p> 2),qedges, and maximum degree∆, then∇(G) ≥q−p+1∆−1 .
Cartesian Product
Definition (G✷H)
V(G✷H) = {(u, v)|u ∈ V(G)andv∈ V(H)}
E(G✷H) = {(u, v)(u′, v′)|u = u′and(v, v′) ∈ E(H)or(u, u′) ∈ E(G)andv= v′}
Min-Yun Lien 23/56
Cartesian Product
Definition (G✷H)
V(G✷H) = {(u, v)|u ∈ V(G)andv∈ V(H)}
E(G✷H) = {(u, v)(u′, v′)|u = u′and(v, v′) ∈ E(H)or(u, u′) ∈ E(G)andv= v′}
Figure :P9✷P11
Decycling number of C
m✷C
nTheorem (Pike, Zou, 2005, SIDMA)
∇(Cm✷Cn) =
⌈3n2⌉ ifm= 4,
⌈3m2 ⌉ ifn= 4,
⌈mn+23 ⌉ otherwise.
Min-Yun Lien 24/56
Decycling number of Path Product P
m✷P
nLower bound
Theorem (Luccio, 1998, IPL)
Ifm, n ≥ 2, then∇(Pm✷Pn) ≥l(m−1)(n−1)+1 3
m .
Decycling number of Path Product P
m✷P
nTheorem (Madelaine and Stewart, 2008, DISC) Table :
In Table, A:∇(Pm✷Pn) = Fm,n, B:∇(Pm✷Pn) ≤ Fm,n+ 1, C:
∇(Pm✷Pn) ≤ Fm,n+ 2, whereFm,n=l(m−1)(n−1)+1 3
m.
Min-Yun Lien 26/56
Decycling number of Path Product P
m✷P
nNew Lower bound Proposition (Observation) Ifm≥ 5andfm,n=(m−1)(n−1)+1
3 is an integer, then each decycling set Sof sizefm,nsatisfies the following two properties:
(1)Scontains exactly one vertex of degree3and contains no vertex of degree 2; and
(2)Sinduces a subgraph ofPm✷Pnwith no edges.
Decycling number of Path Product P
m✷P
nNew Lower bound Proposition (Observation) Ifm≥ 5andfm,n=(m−1)(n−1)+1
3 is an integer, then each decycling set Sof sizefm,nsatisfies the following two properties:
(1)Scontains exactly one vertex of degree3and contains no vertex of degree 2; and
(2)Sinduces a subgraph ofPm✷Pnwith no edges.
Theorem (Lien, Fu, Shih, 2014, DMAA) Ifm≥ 5,mnis even andfm,nis an integer, then
∇(Pm✷Pn) ≥ fm,n+ 1 = Fm,n+ 1,wherefm,n= (m−1)(n−1)+1
3 .
Min-Yun Lien 27/56
Decycling number of Path Product P
m✷P
nProof.
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Min-Yun Lien 28/56
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Min-Yun Lien 28/56
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Letvi,1∈ S, 2 ≤ i ≤ ⌊m2⌋.
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Letvi,1∈ S, 2 ≤ i ≤ ⌊m2⌋.
vm−1,2∈ Sandvm−1,3∈ S./
Min-Yun Lien 30/56
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Letvi,1∈ S, 2 ≤ i ≤ ⌊m2⌋.
vm−1,2∈ Sandvm−1,3∈ S./
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Letvi,1∈ S, 2 ≤ i ≤ ⌊m2⌋.
vm−1,2∈ Sandvm−1,3∈ S./ vm−1,2, vm−1,4, · · · , vm−1,n−1∈ S.
Min-Yun Lien 32/56
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Letvi,1∈ S, 2 ≤ i ≤ ⌊m2⌋.
vm−1,2∈ Sandvm−1,3∈ S./ vm−1,2, vm−1,4, · · · , vm−1,n−1∈ S.
Hence,n− 1is even andnis odd.
Decycling number of Path Product P
m✷P
nProof.
Suppose not.
Assume that∇(Pm✷Pn) = fm,n = Fm,nandSis a decycling set with sizefm,n.
By above Proposition, we may letvi,1be the vertex ofSwith degree3where2≤ i ≤m
2
.
Similarly,
vm−3,n−1, vm−5,n−1, · · · , v2,n−1∈ S.
Thus,mis odd, a contradiction.
Min-Yun Lien 33/56
Decycling number of Path Product P
m✷P
nNew Lower bound Corollary
Form≥ 5, ifm≡ 0(mod6) andn≡ 2 (mod3)or (m, n) ≡ (3, 2) (mod6), then∇(Pm✷Pn) ≥ Fm,n+ 1.
Decycling number of Path Product P
m✷P
nNew Lower bound Corollary
Form≥ 5, ifm≡ 0(mod6) andn≡ 2 (mod3)or (m, n) ≡ (3, 2) (mod6), then∇(Pm✷Pn) ≥ Fm,n+ 1.
Theorem
Form≥ 5, if(m, n) ≡ (0, 2), (0, 5), (3, 2), (2, 0), (5, 0), (2, 3) (mod6), then∇(Pm✷Pn) = Fm,n+ 1.
Min-Yun Lien 34/56
Decycling number of Path Product P
m✷P
nNew Lower bound Corollary
Form≥ 5, ifm≡ 0(mod6) andn≡ 2 (mod3)or (m, n) ≡ (3, 2) (mod6), then∇(Pm✷Pn) ≥ Fm,n+ 1.
Theorem
Form≥ 5, if(m, n) ≡ (0, 2), (0, 5), (3, 2), (2, 0), (5, 0), (2, 3) (mod6), then∇(Pm✷Pn) = Fm,n+ 1.
Theorem (Lien, Fu, Shih, 2014, DMAA) Form, n ≥ 6,∇(Pm✷Pn) ≤ Fm,n+ 1.
Decycling number of Digraphs
5
3 1 0
4 2
Min-Yun Lien 35/56
Decycling number of Digraphs
Notation:
Letvbe a vertex in a digraph. The out-neighborhood or successor setN+(v)is{x ∈ V(G) : v → x}.
ForS⊆ V(G),N+(S) = ∪v∈SN+(v).
Decycling number of Digraphs
Notation:
Letvbe a vertex in a digraph. The out-neighborhood or successor setN+(v)is{x ∈ V(G) : v → x}.
ForS⊆ V(G),N+(S) = ∪v∈SN+(v).
Construction Method Lemma
LetSbe a set of vertices in a digraphG. ThenSis a decycling set of Gif and only if we can find a sequence of subsets ofV(G),
S= S0, S1, · · · , St= V(G)such that (1)Si⊆ Si+1; and
(2)N+(Si+1\ Si) ⊆ Sifori= 0, 1, · · · , t − 1.
Min-Yun Lien 36/56
Decycling number of Digraphs
Example 1.
Decycling number of Digraphs
Example 1.
5
3 1 0
4 2
Min-Yun Lien 37/56
Decycling number of Digraphs
Example 1.
5
3 1 0
4
2 0
Decycling number of Digraphs
Example 1.
5
3 1 0
4
2 0 2
Min-Yun Lien 39/56
Decycling number of Digraphs
Example 1.
5
3 1
4 2
There is a directed cycle(1, 3, 5).
Decycling number of Digraphs
Example 2.
5
3 1 0
4 2
Min-Yun Lien 41/56
Decycling number of Digraphs
Example 2.
5
3 1 0
4
2 0
5
S
1Decycling number of Digraphs
Example 2.
5
3 1 0
4
2 0 2
5 3
S
1Min-Yun Lien 43/56
Decycling number of Digraphs
Example 2.
5
3 1 0
4
2 0 2
5 3
S
1S
2Decycling number of Digraphs
Example 2.
5
3 1 0
4
2 0 2
5 3
1 4
S
1S
2Min-Yun Lien 45/56
Decycling number of Digraphs
Example 2.
5
3 1 0
4
2 0 2
5 3
1 4
S
1S
2 3S
We haveS3= V(G).
Decycling number of Digraphs
Example 2.
3 1
4 2
Min-Yun Lien 47/56
de Bruijn Digraphs and Kautz Digraphs
Definition (de Bruijn digraphB(d, n))
V(B(d, n)) = {x1x2· · · xn: xi∈ {0, 1, · · · , d − 1}, 1 ≤ i ≤ n}.
Edge: X = x1x2· · · xn−→ Y = x2x3· · · xnαwhereα ∈ {0, 1, · · · , d − 1}.
de Bruijn Digraphs and Kautz Digraphs
Definition (de Bruijn digraphB(d, n))
V(B(d, n)) = {x1x2· · · xn: xi∈ {0, 1, · · · , d − 1}, 1 ≤ i ≤ n}.
Edge: X = x1x2· · · xn−→ Y = x2x3· · · xnαwhereα ∈ {0, 1, · · · , d − 1}.
Definition (Kautz digraphK(d, n))
V(K(d, n)) = {x1x2· · · xn : xi∈ {0, 1, · · · , d}, 1 ≤ i ≤ nandxi6= xi+1, 1 ≤ i≤ n − 1}.
Edge: X = x1x2· · · xn−→ Y = x2x3· · · xnαwhereα ∈ {0, 1, · · · , d}.
Remark:K(d, n) ⊆ B(d + 1, n).
Min-Yun Lien 48/56
Generalized de Bruijn Digraphs Generalized Kautz Digraphs
Definition (Generalized de Bruijn digraphGB(d, n)) V(GB(d, n)) = {0, 1, · · · , n − 1}.
E(GB(d, n)) = {(x, y)|y ≡ dx + i (mod n), 0 ≤ i ≤ d − 1}.
Generalized de Bruijn Digraphs Generalized Kautz Digraphs
Definition (Generalized de Bruijn digraphGB(d, n)) V(GB(d, n)) = {0, 1, · · · , n − 1}.
E(GB(d, n)) = {(x, y)|y ≡ dx + i (mod n), 0 ≤ i ≤ d − 1}.
Definition (Generalized Kautz digraphGK(d, n)) V(GK(d, n)) = {0, 1, · · · , n − 1}.
E(GK(d, n)) = {(x, y)|y ≡ −dx − i (mod n), 1 ≤ i ≤ d}.
Min-Yun Lien 49/56
Decycling number of Generalized Kautz Digraphs
Theorem (Lien, Kuo and Fu)
∇(G
k(d, n)) ≤
2
9
n + 3t + 1,
where
n ≡ t (mod 36), for d = 2 ,
n
3
+
94t + 6
where
n ≡ t (mod 36), for d = 3, (
12−
d−12d2)n +
d2(d − t + 5) − 2 ,
where
n ≡ t (mod d + 1), for d ≥ 4.
Decycling number of Generalized de Bruijn Digraphs
Theorem (Lien, Kuo and Fu)
∇(G
B(d, n)) ≤
d+12dn + 2(d − 1).
Min-Yun Lien 51/56
Objective Work
For a planar graphG,∇(G)≤ 2ν(G)? (Jones’ conjecture 2002).
For a planar graphG,∇(G)≤ |V(G)|/2? (Albertson and Berman 1979).
For a bipartite planar graphG,∇(G)≤ 3|V(G)|/8? (Albertson and Berman 1979).
We have⌈(m−1)(n−1)+1
3 ⌉ ≤ ∇(Pm✷Pn) ≤ ⌈(m−1)(n−1)+1
3 ⌉ + 1.Find the exact value of∇(Pm✷Pn).
Find the lower bound of directed graphs.
References
Albertson MO, Berman DM (1979) A conjecture on planar graphs, Bondy JA, Murty USR, Graph theory and related topics 357.
H. Chang, H. L. Fu and M. Y. Lien, The decycling number of outerplanar graphs, J. Comb. Optim. 25 (2013)
Erd ¨os P, Saks M, S ´os VT (1986) Maximum induced trees in graphs. J Combin Theory Ser B 41:61-79.
Bau S, Beineke LW, Vandell RC (1998) Decycling snakes. Congr Numer 134:79-87.
Bodlaender HL (1994) On disjoint cycles. Int J Found Comput Sci 5:59-68.
Erd ¨os P, Saks M, S ´os VT (1986) Maximum induced trees in graphs. J Combin Theory Ser B 41:61-79.
Min-Yun Lien 53/56
Festa P, Pardalos PM, Resende MGC (2000) Feedback set problems, Handbook of Combinatorial Optimization, Du D-Z, Pardalos PM, Eds, Kluwer Academic Publishers, Supplement A, pp 209-259.
Kloks T, Lee C-M, Liu J (2002) New algorithms fork-face cover, k-feedback vertex set, andk-disjoint cycles on plane and planar graphs. in Proceedings of the 28th International Workshop on Graph-Theoretic Concepts in Computer Science (WG 2002) Springer-Verlag, 2573:282-295.
M. Y. Lien, H. L. Fu and C. H. Shih, The decycling number of Pm✷Pn, Discrete Math., Alg. and Appl. DOI:
10.1142/S1793830914500335, 2014.
F. R. Madelaine and I. A. Stewart, Improved upper and lower bounds on the feedback vertex numbers of grids and butterflies,
D. A. Pike and Y. Zou, Decycling Cartesian products of two cycles, SIAM J. Discrete Math., 19 (2005) 651-663.
X. Xu, Y. Cao, J-M. Xu and Y. Wu, Feedback numbers of de Bruijn digraphs, Computers and Mathematics with Application, 59 (2010) 716-723.
F. R. Madelaine and I. A. Stewart, Improved upper and lower bounds on the feedback vertex numbers of grids and butterflies, Discrete Math., 308 (2008) 4144-4164.
D. A. Pike and Y. Zou, Decycling Cartesian products of two cycles, SIAM J. Discrete Math., 19 (2005) 651-663.
X. Xu, Y. Cao, J-M. Xu and Y. Wu, Feedback numbers of de Bruijn digraphs, Computers and Mathematics with Application, 59 (2010) 716-723.
Min-Yun Lien 55/56