Note on the Burning Conjecture for Some Graphs

Abstract

Graph burning is a model for describing the spread of influence in social networks and the burning number is a parameter used to describe the speed of information spread. In 2016, Bonato proposed a graph burning conjecture: For any connected graph G with order n , the burning number b( G ) n . In this paper, we confirm the burning conjecture for octopus graph and bicyclic graph.

Share and Cite:

Zhu, Q. and Li, Y. (2025) Note on the Burning Conjecture for Some Graphs. Open Journal of Applied Sciences, 15, 1157-1167. doi: 10.4236/ojapps.2025.155080.

1. Introduction

Graph burning is a model that describes the spread of social contagion on social networks such as Facebook or Twitter. We use Bondy and Murty [1] for the notation and terminology, burning process is defined as follows. Given a finite and simple graph G , vertices may be either burned or unburned throughout the process. Initially, at time t=0 , all vertices are unburned. At each time t1 , one new unburned vertex is chosen to burn, if such a vertex is available, we call such a chosen vertex a fire source. If a vertex is burned, then it remains in that state until the end of the process. Once a vertex is burned at time t , at time t+1 each of its unburned neighbors becomes burned. The process ends when all vertices of G are burned.

Note that the burning process on G may be highly dependent on the choice of fire sources, the strategic choice of sources is critical when minimizing the length of the process. In [2], Bonato introduced the burning number of graph G , denoted by b( G ) , is the minimum steps to burn graph G . The fire sources x 1 ,, x k that are chosen over time on graph G are referred to as a burning sequence ( x 1 ,, x k ) of G and call the shortest burning sequence optimal. Clearly, optimal burning sequences have length b( G ) .

In 2016, Bonato et al proposed the burning conjecture:

Conjecture 1.1 [3]. For a connected graph G of order n , b( G ) n .

Later, the conjecture been comfirm for some graph classes, such as spiders [4], path forests [5], caterpillars [6] [7], theta graph [8], fence graph [9], generalized Petersen graph [10], the Cartesian product of paths [11], Q graph [12], binary tree [13] and trees without degree-2 vertices [14].

Motivated by these, we put forward on the paths and circles. We first denote octopus graph G which obtained from comet graph C l i ,2 for 1im by identifing the tail of C l i ,2 at v 0 , clearly, d( v 0 )=m . All degree-3 vertices but v 0 denoted by u i , we call C l i ,2 an arm of Octopus and l i is the length of arm C l i ,2 (see Figure 1). Here the comet C r,s is a graph which obtained by the end of path P r with the center of star graph K 1,s .

Figure 1. Octopus graph G.

Another class of graph named t tail bicyclic graph. If t=1 , call single tail bicyclic graph which obtained by joining the center vertex of the bicyclic graphs with one vertex of the path P a 1 +1 denoted by G v 0 ( g 1 , g 2 , a 1 ) . If t=2 , call double tail bicyclic graph which obtained by joining the center vertex of the bicyclic graph with a vertex of path P a 1 +1 and path P a 2 +1 denoted by G v 0 ( g 1 , g 2 , a 1 , a 2 ) (see Figure 2). They all only have a vertex v 0 with d( v 0 )>2 , without loss of generality, we suppose g 1 g 2 .

Figure 2. t tail bicyclic graph G v 0 ( g 1 , g 2 , a 1 ) and G v 0 ( g 1 , g 2 , a 1 , a 2 ) ( t=1,2 ).

In this paper, we first confirm the burning conjecture for these 3 kind of graphs, next we discuss the burning number of single tail bicyclic graph G v 0 ( g 1 , g 2 , a 1 ) and double tail bicyclic graph G v 0 ( g 1 , g 2 , a 1 , a 2 ) .

2. Primarilies

Lemma 2.1 [2] If ( x 1 , x 2 ,, x k ) is a sequence of nodes in a graph G such that N k1 [ x 1 ] N k2 [ x 2 ] N 1 [ x k1 ] N 0 [ x k ]=V( G ) , then b( G )k .

Lemma 2.2 [3] For a path P n or a cycle C n on n nodes, we have b( P n )=b( C n ) n .

Lemma 2.3 [3] For a graph G , b( G )=min{ b( T ):TisaspanningtreeofG } .

Lemma 2.4 [3] For any graph G with radius r and diameter d , we have that ( d+1 ) 1 2 b( G )r+1 .

Lemma 2.5 [4] The burning number of a spider graph G of order n satisfies b( G ) n .

Lemma 2.6 [4] If G is a path-forest of order n with t1 components, then b( G ) n 2 +t .

Lemma 2.7 [5] Let G= P a 1 P a 2 with a 1 a 2 1 and J( t )={ ( t 2 2,2 ) } for integer t2 . Then

b( G )={ a 1 + a 2 +1, If( a 1 , a 2 )J( t ); a 1 + a 2 , Otherwise.

Lemma 2.8 [5] Let G= P a 1 P a 2 P a 3 with a 1 a 2 a 3 1 . Then

b( G )={ a 1 + a 2 + a 3 +1, If( a 1 , a 2 , a 3 ) J 1 J 2 J 3 J 4 J 5 ; a 1 + a 2 + a 3 , Otherwise.

Let J i for 1i5 satisfy the following conditions.

D 1 ={ ( 2,2 ) },

D 2 ={ ( 3,2 ) },

D 3 ={ ( 1,1 ),( 3,3 ),( 4,2 ),( 5,5 ) },

D 4 ={ ( 2,1 ),( 4,1 ),( 4,3 ),( 4,4 ),( 6,1 ),( 6,4 ),( 6,5 ),( 6,6 ),( 7,7 ),( 8,4 ),( 8,6 ),( 10,4 ) },

D 5 ={ ( 11,10,4 ) ,( 13,11,1 ),( 11,11,3 ),( 22,13,1 ),( 19,13,4 ),( 17,13,6 ),( 15,13,8 ), ( 13,13,10 ),( 17,15,4 ),( 15,15,6 ),( 30,15,4 ),( 28,15,6 ),( 26,15,8 ),( 19,15,15 ), ( 28,17,4 ),( 26,17,6 ),( 17,17,15 ),( 26,19,4 ),( 43,17,4 ),( 41,17,6 ),( 30,17,17 ), ( 41,19,4 ),( 30,30,4 ), ( 58,19,4 ) },

J 1 ={ ( a 1 , a 2 , a 3 ):( a 2 , a 3 ) D 1 , a 1 + a 2 + a 3 = t 2 3forintegert },

J 2 ={ ( a 1 , a 2 , a 3 ):( a 2 , a 3 ) D 1 D 2 , a 1 + a 2 + a 3 = t 2 2forintegert },

J 3 ={ ( a 1 , a 2 , a 3 ):( a 2 , a 3 ) i=1 3 D i , a 1 + a 2 + a 3 = t 2 1forintegert },

J 4 ={ ( a 1 , a 2 , a 3 ): a 3 =2or( a 2 , a 3 ) i=1 4 D i , a 1 + a 2 + a 3 = t 2 forintegert },

J 5 = D 5 { 11,11,2 }.

3. Main Results

In this section, we first confirm the burning conjecture for octopus graph G , single tail bicyclic graph G v 0 ( g 1 , g 2 , a 1 ) and double tail bicyclic graph G v 0 ( g 1 , g 2 , a 1 , a 2 ) .

Theorem 3.1 Let G be a octopus graph with order n . Then b( G ) n

Proof. Let G be a octopus graph with order n . Without loss of generality, suppose n= q 2 +p for 1p2q+1 . The neighbors of v 0 are v 1 , v 2 ,, v i respectively and the length of longest arm is l . If q=2 , it’s clearly that b( G ) n . Consider q3 , next we distinguish 3 cases to complete the proof.

Case 1 If lq .

It’s clearly that radius of G is rq , by lemma 2.4, we have b( G ) n .

Case 2 If q+1l2q1 .

Consider the structure of the arm, we will discuss two cases.

Subcase 2.1 The arms of G have the same structure.

If each arm has the same length, then G has q arms of length q+1 . First we set the x 1 on v 1 , then C l 1 ,2 N q [ x 1 ] . We denoted the part of G\ N q [ x 1 ] is R= R 1 R 2 R q1 , clearly, the height of R i ( 1iq1 ) is 2 and each R i contains u i . Now suppose ( x 2 , x 3 ,, x q ) is a burning sequence of R and let x i+1 = u i for 1iq1 , then we have V( G ) N q [ x 1 ] N q1 [ x 2 ] N q2 [ x 3 ] N 1 [ x q ] . By lemma 2.1, we have b( G ) n =q+1 .

Subcase 2.2 The arms of G doesn’t have the same structure.

It’s clearly that G doesn’t have q arms of length q+1 . We set x 1 on v 0 , if l i q , then C l i ,2 N q [ x 1 ] . We denoted the part of G\ N q [ x 1 ] is R 1 R 2 R i ( i1 ) . Since the arm of octopus G has length at most 2q1 , then each R i has length and order at most q2 and q respectively. Next we discuss the burning number of G by the number of i compontent.

When i q+1 2 . We remove a pendant vertex to another pendant vertex such that R i become a path P i , P= P 1 P 2 P i ( i1 ) , it’s clearly that b( R )b( P ) . Next we denoted w k is ssthe center of P k , for each k=1,2,,i , we can by neighborhood N qk [ w k ] cover the P k . since 2( qk )+12 q+1 2 +1=q , thus b( R )b( P )q . By lemma 2.1, we have b( G ) n =q+1 .

When i= q+2 2 , q is even. For each k=1,2,,i1 , every R k can be covered by N qk [ w k ] . For R i , it has the lenght at most q2 , if R i only have 2 isolated vertex, we set w i at u i , otherwise w i is the center of R i , 2( qi )=2( q q+2 2 )=q2 , thus R i N qi [ w i ] . By lemma 2.1, we have b( G ) n =q+1 .

When i q+4 2 . Since lq+1 and G doesn’t have q arms of length q+1 , then there must exist an arm C l i ,2 with the length | l i |q . C l i ,2 { v 0 } contain at least 3 vertices, then N q [ x 1 ] contains at least iq+4 vertices, thus | R |niq13q( q+2i )3 . If i=q+1 , then | R |q3i=q+1 , the number of vertices is smaller than the number of branches, a contradiction. If i=q , then | R |2q3i=q( q3 ) , a satisfaction.

So we only consider q+4 2 iq , removing a pendant vertex to another pendant vertex doesn’t change the number of vertices of R , thus V( R )=V( P ) . By b( R )b( P ) and lemma 2.6, we have

b( R )b( P ) V( R ) 2i +i q( q+2i )3 2i +i

Let f( i )= q( q+2i )3 2i +i , because of the properties of the function, the maximum is attained at the q or q+4 2 .

Suppose i=q , we have

b( R ) V( R ) 2i +i= q( q+2i )3 2i +i= 2q3 2q +qq

Suppose i= q+4 2 , we have

b( R ) V( R ) 2i +i= q( q+2i )3 2i +i = q( q+2 q+4 2 3 ) 2 q+4 2 + q+4 2

when q is even, then q+4 2 = q+4 2 , we have b( R ) q( q+2 q+4 2 3 ) 2 q+4 2 + q+4 2 = q 2 2q+3 q+4 + q+4 2 , since 2q+3 q+4 ( 1,2 ) , then q 2 2q+3 q+4 = q 2 2 , thus we have b( R ) q 2 2+ q+4 2 =q .

When q is odd, then q+4 2 = q+4 2 , we have b( R ) q( q+2 q+4 2 3 ) 2 q+4 2 + q+4 2 = q 2 1 + q+3 2 , since q 2 1 = q1 2 1 , thus we have b( R ) q1 2 1+ q+3 2 =q .

We know G= N q [ x 1 ]R , suppose ( z 1 , z 2 ,, z q ) is a burning sequence of R . Next let x 1 = v 0 and x i+1 = z i for 1iq , it is clear that V( G ) N q [ x 1 ] N q1 [ x 2 ] N q2 [ x 3 ] N 0 [ x q+1 ] , by lemma 2.1 we have b( G )q+1 .

Case 3 If l2q .

When l2q , we proceed with contradiction, assume G c is the minimal counterexample of octopus graph with order n . This means

b( G c )> | b( G c ) | =q+1 . Suppose the length of longest arm C l i ,2 of G c is l , we have w 1 C l i ,2 such that d( v 0 , w 1 )=l . Since l2q , we have w 0 C l i ,2 satisfied d( w 0 , w 1 )=q1 . We set x 1 at w 0 , Clearly, N q [ w 0 ] can burn 2q+1 vertices, We denoted G 1 =G\ N q [ w 0 ] where v 0 G 1 , then | G 1 | = |G\ N q [ w 0 ] | q 2 +2q+1( 2q+1 ) q 2 . Consider G c is a counterexample of octopus with minimal number of vertices, we have b( G 1 ) | G 1 | q , then ( z 1 , z 2 ,, z q ) is a burning sequence of G 1 . Next let x 1 = w 0 , x i+1 = z i for 1iq , it is clear that V( G c ) N q [ x 1 ] N q1 [ x 2 ] N q2 [ x 3 ] N 0 [ x q+1 ] . By lemma 2.1 we have b( G c )q+1 . This contradicts to the fact b( G c )>q+1 . Thus, we have b( G ) n . □

Theorem 3.2 If G= G v 0 ( g 1 , g 2 , a 1 ) is a single tail bicyclic graph with order n , then n+ 21 4 3 2 b( G ) n .

Proof. We take a edge e i from C g i (i = 1, 2), then we can derive

G v 0 ( g 1 , g 2 , a 1 ) = G v 0 ( g 1 , g 2 , a 1 ) i=1 2 e i is a spider graph. By lemma 2.3 and lemma 2.5, we have b( G v 0 ( g 1 , g 2 , a 1 ) b( G v 0 ( g 1 , g 2 , a 1 ) ) n .

Next, we prove the lower bound, suppose b( G )=k and ( x 1 , x 2 ,, x k ) is an optimal burning sequence of G , we set x 1 on v 0 to contains more vertices, then, N k1 [ x 1 ]5( k1 )+1 , combine with the fact that | N ki [ x i ] |2( ki )+1 for 2ik , we get

| N k1 [ x 1 ] |+ i=2 k ( 2( ki )+1 )( 5( k1 )+1 )+( 2( k2 )+1 )++1 =( 5k4 )+( 2k3 )++1 = k 2 +3k3.

by k 2 +3k3n , we have k n+ 21 4 3 2 . □

Theorem 3.3 If G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) is a double tail bicyclic graph with order n , then n+8 2b( G ) n .

Proof. We take a edge e i from C g i ( i=1,2 ), then we can derive G v 0 ( g 1 , g 2 , a 1 , a 2 ) = G v 0 ( g 1 , g 2 , a 1 , a 2 ) i=1 2 e i is a spider graph. by lemma 2.3 and lemma 2.5, we have b( G v 0 ( g 1 , g 2 , a 1 , a 2 ) )b( G v 0 ( g 1 , g 2 , a 1 , a 2 ) ) n .

Next, we prove the lower bound, suppose b( G )=k and ( x 1 , x 2 ,, x k ) is an optimal burning sequence of G . We set x 1 on v 0 to contains more vertices, then, N k1 [ x 1 ]6( k1 )+1 , combine with the fact that | N ki [ x i ] |2( ki )+1 for 2ik , we have

| N k1 [ x 1 ] |+ i=2 k ( 2( ki )+1 )( 6( k1 )+1 )+( 2( k2 )+1 )++1 =( 6k5 )+( 2k3 )++1 = k 2 +4k4.

by k 2 +4k4n , we have k n+8 2 . □

We following discuss the burning number of G v 0 ( g 1 , g 2 , a 1 ) and G v 0 ( g 1 , g 2 , a 1 , a 2 ) .

Consider G= G v 0 ( g 1 , g 2 , a 1 ) , by Theorem 3.2, we have

Corollary 3.4 If G= G v 0 ( g 1 , g 2 , a 1 ) is a single tail bicyclic graph with order q 2 +p for 1p2q+1 , then q1b( G )q+1 .

Next we discuss the graph G v 0 ( g 1 , g 2 , a 1 ) with burning number q+1 .

Theorem 3.5 Let G= G v 0 ( g 1 , g 2 , a 1 ) be a single tail bicyclic graph with order q 2 +p . If 3q2p2q+1 , then b( G )=q+1

Proof. As we know from the previous, if b( G )=q , then it can contains at most q 2 +3q2 vertices of G . Since 3q2p2q+1 , then we have b( G )q+1 , combine with corollary 3.4, we have b( G )=q+1 . □

Theorem 3.6 Let G= G v 0 ( g 1 , g 2 , a 1 ) be a single tail bicyclic graph with order q 2 +p for 1p2q+1 . If g 1 q 2 or a 1 q 2 , then b( G )=q+1 .

Proof. We discuss two cases to complete the proof.

Case 1 If g 1 q 2 .

In this case, let H= C g 1 +1 , | H | q 2 +1 , by lemma 2.2, then b( H )q+1 . If g 2 q and a 1 q , then b( G )=b( H )q+1 . If g 2 q or a 1 q , then b( G )b( H )q+1 . Thus we have b( G )q+1 , combine with corollary 3.4, we have b( G )=q+1 .

Case 2 If a 1 q 2 .

In this case, let H= P a 1 +1 , | P a 1 +1 | q 2 +1 , by lemma 2.2, b( H )q+1 . H is a subgraph of G , when g 1 2 q , we have b( G )=b( H )q+1 , when g 1 2 >q , we have b( G )>b( H )q+1 . Thus we have b( G )q+1 , combine with corollary 3.4, we have b( G )=q+1 . □

Theorem 3.7 Let G= G v 0 ( g 1 , g 2 , a 1 ) be a single tail bicyclic graph with order q 2 +p for 1p2q+1 . If g 1 2 + a 1 q 2 or g 1 2 + g 2 2 q 2 , then

b( G )=q+1 . □

Proof. According to the definition of diameter, d( G )=max{ g 1 2 + a 1 , g 1 2 + g 2 2 } , it’s clearly that d( G ) q 2 , by lemma 2.4, we have b( G ) d( G )+1 q 2 +1 =q+1 , combine with corollary 3.4, we have b( G )=q+1 .

Theorem 3.8 Let G= G v 0 ( g 1 , g 2 , a 1 ) be a single tail bicyclic graph with order q 2 +p for 1p2q+1 . If 2q2 g 1 q 2 +1 , 2q2 g 2 q 2 +1 , q1 a 1 q 2 +1

1) If min{ g 1 , g 2 , a 1 }0 , { g 1 , g 2 , a 1 } J 1 J 2 J 3 J 4 J 5 , then b( G )=q+1 .

2) If min{ g 1 , g 2 , a 1 }=0 , { g 1 , g 2 , a 1 }J( t ) , then b( G )=q+1 .

where g 1 = g 1 ( 2q2 ), g 2 = g 2 ( 2q2 ), a 1 = a 1 ( 2q2 ) .

Proof. 1) If min{ g 1 , g 2 , a 1 }0 , suppose ( x 1 , x 2 ,, x q ) is an optimal burning sequence of G , we set x 1 at v 0 , then | N q1 [ x 1 ] |=5q4 , H=G N q1 [ x 1 ]= P g 1 P g 2 P a 1 . Since { g 1 , g 2 , a 1 } J 1 J 2 J 3 J 4 J 5 , by lemma 2.8, we have b( H )= g 1 + g 2 + a 1 +1 n( 5q4 ) +1 ( q1 ) 2 +1q , then b( G )=q+1 , a contradiction, thus we have b( G )q+1 , combine with corollary 3.4, we have b( G )=q+1 .

2) If min{ g 1 , g 2 , a 1 }=0 , suppose ( x 1 , x 2 ,, x q ) is an optimal burning sequence of G , we set x 1 at v 0 , then | N q1 [ x 1 ] |=5q4 ,

H=G N q1 [ x 1 ]= P g 1 P g 2 P a 1 . Since { g 1 , g 2 , a 1 }J( t ) , by lemma 2.7, we have b( H ) n( 5q4 ) +1 ( q1 ) 2 +1q , then b( G )=q+1 , a contradiction, thus we have b( G )q+1 . combine with corollary 3.4, we have b( G )=q+1 . □

Consider G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) , by Theorem 3.3, we have

Corollary 3.9 If G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) is a double tail bicyclic graph with order q 2 +p for 1p2q+1 , then q1b( G )q+1 .

Lemma 3.10 If G is disconnected with connected components G 1 , G 2 ,, G i , each G i contains no isolated vertices, then b( G )b( G 1 )+b( G 2 )++b( G i )i+1 .

Proof. For each G j ( 1ji ) , we suppose X j =( x 1 ( j ) , x 2 ( j ) ,, x b( G j ) ( j ) ) is an optimal burning sequence, Clearly G also has a burning sequence. We claim X=( x 1 ( 1 ) ,, x b( G 1 )1 ( 1 ) , x 1 ( 2 ) ,, x b( G 2 )1 ( 2 ) ,, x 1 ( i ) ,, x b( G i )1 ( i ) , x b( G i ) ( i ) ) is a burning sequence of G .

For each j{ 1,2,,i1 } , we burn X\{ x b( G j ) ( j ) } in order. Before we burn x 1 ( i ) , C i1 need at most 2 rounds to burn complete. Since G i contains no isolated vertices, then b( G i )2 . Therefore, when G i is burned, G i1 has enough time to burn completely, thus all the G j can be burned completely. since X is a valid burning sequence of G , thus we have

b( G )| X |=( b( G 1 )1 )++( b( G i )1 )+b( G i )=b( G 1 )+b( G 2 )++b( G i )i+1 .

Next we discuss G v 0 ( g 1 , g 2 , a 1 , a 2 ) with burning number q+1 .

Theorem 3.11 Let G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) be a double tail bicyclic graph with order q 2 +p for 1p2q+1 . If q=2,p=5 , then b( G )=q+1 .

Proof. If q=2 , then n[ 5,9] ,p [1,5 ] . If b( G )=q , it can contain at most q 2 +4q4 vertices. When p4q3 , then p=5 , thus b( G )q+1 . Combine with corollary 3.9, we have b( G )=q+1 . □

Theorem 3.12 Let G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) be a double tail bicyclic graph with order q 2 +p for 1p2q+1 . If g 1 q 2 or a 1 + a 2 q 2 , then b( G )=q+1 .

Proof. We discuss two cases to complete the proof.

Case 1 If g 1 q 2 .

In this case, let H= C g 1 +1 , | H | q 2 +1 , by lemma 2.2, then b( H )q+1 . Because of the symmetry of the circle, we set x 1 at v 0 . If max{ g 2 2 , a 1 , a 2 }q , then b( G )=b( H )q+1 , if max{ g 2 2 , a 1 , a 2 }>q , then b( G )>b( H )q+1 . Thus we can derive b( G )q+1 , combine with corollary 3.9, we have b( G )=q+1

Case 2 If a 1 + a 2 q 2 .

In this case, let H= P a 1 + a 2 +1 , | P a 1 + a 2 +1 | q 2 +1 , by lemma 2.2, b( H )q+1 . H is a subgraph of G , similar to case 1, we have b( G )b( H )q+1 , combine with corollary 3.9, we have b( G )=q+1 . □

Theorem 3.13 Let G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) be a double tail bicyclic graph with order q 2 +p for 1p2q+1 . If g i 2 + a i q 2 or g 1 2 + g 2 2 q 2 , then b( G )=q+1 .

Proof. According to the definition of diameter,

d( G )=max{ g i 2 + a i , g 1 2 + g 2 2 } , it’s clearly that d( G ) q 2 +1 , by lemma 2.4, we have b( G ) d( G )+1 q 2 +1 =q+1 , combine with corollary 3.9, we have b( G )=q+1 . □

Theorem 3.14 Let G= G v 0 ( g 1 , g 2 , a 1 , a 2 ) be a double tail bicyclic graph with order q 2 +p for 1p2q+1 . If 2q2 g 1 q 2 +1 , 2q2 g 2 q 2 +1 , q1 a 1 q 2 +1 , q1 a 2 q 2 +1

1) If min{ g 1 , g 2 , a 1 , a 2 }0 , g 1 1 , g 2 1 , a 1 1 , a 2 1 , g 1 + g 2 + a 1 + a 2 q+3 , then b( G )=q+1 .

2) If min{ g 1 , g 2 , a 1 , a 2 }=0 , { g 1 , g 2 , a 1 , a 2 } J 1 J 2 J 3 J 4 J 5 or { g 1 , g 2 , a 1 , a 2 }J( t ) , then b( G )=q+1 .

where g 1 = g 1 ( 2q2 ), g 2 = g 2 ( 2q2 ), a 1 = a 1 ( 2q2 ), a 2 = a 2 ( 2q2 )

Proof. 1) If min{ g 1 , g 2 , a 1 , a 2 }0 , suppose ( x 1 , x 2 ,, x q ) is an optimal burning sequence of G , we set x 1 at v o , then | N q1 [ x 1 ] |=6q5 , H=G N q1 [ x 1 ]= P g 1 P g 2 P a 1 P a 2 , Since g 1 1 , g 2 1 , a 1 1 , a 2 1 , g 1 + g 2 + a 1 + a 2 q+3 , by lemma 3.10, we have b( H ) g 1 + g 2 + a 1 + a 2 4+1q , then b( G )=q+1 , a contradiction, thus we have b( G )q+1 . combine with corollary 3.9, we have b( G )=q+1 .

2) If min{ g 1 , g 2 , a 1 , a 2 }=0 , suppose ( x 1 , x 2 ,, x q ) is an optimal burning sequence of G , we set x 1 at v o , then | N q1 [ x 1 ] |=6q5 , H=G N q1 [ x 1 ]

When { g 1 , g 2 , a 1 , a 2 } J 1 J 2 J 3 J 4 J 5 , by lemma 2.8, then

b( H ) n( 6q5 ) +1 ( q2 ) 2 +2 +1q , then b( G )=q+1 , a contradiction, thus we have b( G )q+1 . combine with corollary 3.9, we have b( G )=q+1 .

When { g 1 , g 2 , a 1 , a 2 }J( t ) , by lemma 2.7, we have

b( H )= n( 6q5 ) +1 ( q2 ) 2 +2 +1q , then b( G )=q+1 , a contradiction, thus we have b( G )q+1 . combine with corollary 3.9, we have b( G )=q+1 . □

4. Conclusion

In this paper, we put forward on the unions of paths and circlies and confirm the burning conjecture for octopus graphs and t tail bicyclic graph ( t=1,2 ), we also discuss the single tail bicyclic graph and double tail bicyclic graph with the burning number q+1 . The burning conjecture has been a topic of concern which be useful for information dissemination. Our study is meaningful and next we focus on the burning number for these graph. Besides we will extend graph burning to hypergraph and achieve more results.

Acknowledgements

The authors would like to thank anonymous reviewers for their valuable comments and suggestions to improve the quality of this article. This research was supported by Qinghai Minzu University Innovative Project (No. 07M2024006), NSFC No. 12371352 and Qinghai provincial basic research project No. 2022-ZJ-753.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.

References

[1] Bondy, J.A. and Murty, U.S.R. (2008) Graph Theory. GTM 244, Springer.
[2] Bonato, A., Janssen, J. and Roshanbin, E. (2014) Burning a Graph as a Model of Social Contagion. In: Bonato, A., Graham, F. and Prałat, P., Eds., Lecture Notes in Computer Science, Springer International Publishing, 13-22.[CrossRef]
[3] Bonato, A., Janssen, J. and Roshanbin, E. (2016) How to Burn a Graph. Internet Mathematics, 12, 85-100.[CrossRef]
[4] Bonato, A. and Lidbetter, T. (2019) Bounds on the Burning Numbers of Spiders and Path-Forests. Theoretical Computer Science, 794, 12-19.[CrossRef]
[5] Liu, H., Hu, X. and Hu, X. (2021) Burning Numbers of Path Forests and Spiders. Bulletin of the Malaysian Mathematical Sciences Society, 44, 661-681.[CrossRef]
[6] Liu, H., Hu, X. and Hu, X. (2020) Burning Number of Caterpillars. Discrete Applied Mathematics, 284, 332-340.[CrossRef]
[7] Hiller, M., Koster, A.M.C.A. and Triesch, E. (2021) On the Burning Number of P-Caterpillars. In: Gentile, C., Stecca, G. and Ventura, P., Eds., AIRO Springer Series, Springer International Publishing, 145-156.[CrossRef]
[8] Liu, H., Zhang, R. and Hu, X. (2019) Burning Number of Theta Graphs. Applied Mathematics and Computation, 361, 246-257.[CrossRef]
[9] Bonato, A., English, S., Kay, B. and Moghbel, D. (2021) Improved Bounds for Burning Fence Graphs. Graphs and Combinatorics, 37, 2761-2773.[CrossRef]
[10] Sim, K.A., Tan, T.S. and Wong, K.B. (2018) On the Burning Number of Generalized Petersen Graphs. Bulletin of the Malaysian Mathematical Sciences Society, 41, 1657-1670.[CrossRef]
[11] Mitsche, D., Prałat, P. and Roshanbin, E. (2018) Burning Number of Graph Products. Theoretical Computer Science, 746, 124-135.[CrossRef]
[12] Li, Y., Wu, J., Qin, X. and Wei, L. (2024) Characterization of $ Q $ Graph by the Burning Number. AIMS Mathematics, 9, 4281-4293.[CrossRef]
[13] Das, Sandip, Islam, S.S., Mitra, R.M. and Paul, S. (2023) Burning a Binary Tree and Its Generalization. arxiv:2308.02825.
[14] Murakami, Y. (2024) The Burning Number Conjecture Is True for Trees without Degree-2 Vertices. Graphs and Combinatorics, 40, Article No. 82.[CrossRef]

Copyright © 2026 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.