Next Article in Journal
A Scaled Dai–Yuan Projection-Based Conjugate Gradient Method for Solving Monotone Equations with Applications
Next Article in Special Issue
Graph Coloring via Clique Search with Symmetry Breaking
Previous Article in Journal
A New Extension to the Intuitionistic Fuzzy Metric-like Spaces
Previous Article in Special Issue
Some Results on the Erdős–Faber–Lovász Conjecture
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Note on the Lagrangian of Linear 3-Uniform Hypergraphs

1
School of Mathematics, Hunan University, Changsha 410082, China
2
CHP-LCOCS, School of Mathematics and Statistics, Hunan Normal University, Changsha 410081, China
*
Author to whom correspondence should be addressed.
Submission received: 11 June 2022 / Revised: 2 July 2022 / Accepted: 5 July 2022 / Published: 7 July 2022
(This article belongs to the Special Issue Symmetry in Graph and Hypergraph Theory)

Abstract

:
Lots of symmetric properties are well-explored and analyzed in extremal graph theory, such as the well-known symmetrization operation in the Turán problem and the high symmetric in the extremal graphs. This paper is devoted to studying the Lagrangian of hypergraphs, which connects to a very symmetric function—the Lagrangian function. Given an r-uniform hypergraph F, the Lagrangian density π λ ( F ) is the limit supremum of r ! λ ( G ) over all F-free G, where λ ( G ) is the Lagrangian of G. An r-uniform hypergraph F is called λ -perfect if π λ ( F ) equals r ! λ ( K v ( F ) 1 r ) . Yan and Peng conjectured that: for integer r 3 , there exists n 0 ( r ) such that if G and H are two λ -perfect r-graphs with | V ( G ) | and | V ( H ) | no less than n 0 ( r ) , then the disjoint union of G and H is λ -perfect. Let S t denote a 3-uniform hypergraph with t edges { e 1 , , e t } satisfying that e i e j = { v } for all 1 i < j t . In this paper, we show that the conjecture holds for G = S 2 and H = S t for all t 62 . Moreover, our result yields a class of Turán densities of 3-uniform hypergraphs. In our proof, we use some new techniques to study Lagrangian density problems; using a result of Sidorenko to find subgraphs, and a result of Talbot to upper bound the Lagrangian of a hypergraph.

1. Introduction

Symmetry is a major characteristic of mathematical beauty, and it is found in many branches of mathematics. A number of symmetric properties are widely studied, analyzed and applied in graph theory. For example, a well-known symmetrization operation is widely applied in the study of the Turán type problems in extremal graph theory, and the Lagrangian of hypergraphs concerned in this paper connects to a very symmetric function–the Lagrangian function (see [1,2] for surveys).
For a finite set X and an integer r > 0 , let X ( r ) = { A X : | A | = r } . An r-uniform hypergraph (r-graph) G on the vertex set V ( G ) , is a subset of V ( G ) ( r ) . We simply denote G as the edge set of G. We call a 2-graph a simple graph. Denote G ¯ = V ( G ) ( r ) G . Let S V ( G ) , denote G S = { e G : e S = } and G [ S ] = { e G : e S } . An edge e = { v 1 , v 2 , , v r } will be simply denoted by v 1 v 2 v r . Let K n r = X ( r ) denote the complete r-graph on vertex set X with | X | = n . Denote [ n ] = { 1 , 2 , , n } . Given an r-graph G on vertex set [ n ] , define the Lagrangian function of G as
w ( G , x ) = e G i e x i ,
where x [ 0 , ) n . w ( G , x ) can be interpreted as the density of a blow-up of G divide r ! . Define the Lagrangian of G as λ ( G ) = max x Δ w ( G , x ) , where
Δ = x R n : i = 1 n x i = 1 , x i 0 .
We call a weighting x a feasible weighting if x Δ . We call a feasible weighting x optimal if w ( G , x ) = λ ( G ) . Given r-graphs G and F, G is said to be F-free if G contains no copy of F. The Lagrangian density  π λ ( F ) of F is defined to be
π λ ( F ) = r ! sup { λ ( G ) : G is F - free } .
The idea of continuous optimization is widely used not only in mathematics, but also in other disciplines (see [3,4] and so on). The hypergraph Lagrangian method, first introduced by Zykov [5] in 1949, is such a continuous optimization method that is helpful to solve the extremal problems. One of the earliest applications of the hypergraph Lagrangian method was applied by Motzkin and Straus [6] in 1965 to establish the connection (See Theorem 1) between the Lagrangian of a simple graph and its maximum clique number. A surprising application is that in the 1980’s, Frankl and Rödl [7] used it to disprove a famous conjecture of Erdos. For more developments of the Lagrangian theory of hypergraphs see [8,9]. Actually, determining the Lagrangian density of general r-graphs when r 3 is interesting in itself. However, there are very few known results in Lagrangian density problems. We list some of the relevant results known so far as follows.
Since K n 1 r is F-free for an n-vertex r-graph F, then
π λ ( F ) r ! λ ( K n 1 r ) .
We call an n-vertex r-graph F λ -perfect if π λ ( F ) = λ ( K n 1 r ) (We will show the value of λ ( K n r ) in Fact 4).
Motzkin and Straus [6] showed that every 2-graph K t 2 is λ -perfect. Next, we turn our attention to the hypergraphs. Let T be a tree or a forest that satisfies Erdos and Sós’ conjecture, and let F be an r-graph obtained by joining r 2 fixed vertices into every edge of T. Sidorenko [10] proved that F is λ -perfect for large tree. Let H r be the r-graph with edge set { v 1 v r 1 v r , v 1 v r 1 v r + 1 } . Sidorenko [9] showed that H r is λ -perfect for r = 3 and 4. Let M t r be the r-uniform matching with t pairwise disjoint edges. Let S t denote a 3-uniform hypergraph with t edges { e 1 , , e t } satisfying that e i e j = { v } for all 1 i < j t . Hefetz and Keevash [11] showed that M 2 3 is λ -perfect. The authors [12] proved that M t 3 and S t 4 are λ -perfect. A result given by Bene Watts, Norin and Yepremyan [13] suggested that M 2 r is not λ -perfect for r 4 . It is interesting to study the λ -perfect r-graphs. More results yielding λ -perfect r-graphs are in the papers [10,14,15,16,17,18,19,20,21,22]. It is worth mentioning that Yan and Peng [23] recently proved that the Lagrangian density of a 3-graph is an irrational number, and independently, Wu [24] showed that the Lagrangian density of M 3 4 is an irrational number. These two results give a positive answer to the question posed by Baber and Talbot [25]: whether there is an irrational Turán density of a single hypergraph. For more relevant Hypergraph Lagrangian results, one may refer to [10,17,22,26,27,28,29,30,31,32,33,34,35,36,37,38].
We call a graph linear if any two edges of it share at most one vertex in common. Our original motivation is to seek some new tools to study the Lagrangian density of linear 3-graphs, and to give brief proofs. Denote the disjoint union of two r-graphs G and H as G H . In 2019, Yan and Peng [22] posed an interesting conjecture.
Conjecture 1 
(Yan and Peng [22]).
(i) 
For an integer r 3 , there exists n 0 ( r ) such that a linear r-graph on at least n 0 ( r ) vertices is λ -perfect.
(ii) 
For an integer r 3 , there exists n 0 ( r ) such that if G and H are two λ -perfect r-graphs with v ( G ) n 0 ( r ) and v ( H ) n 0 ( r ) , then G H is λ -perfect.
The other motivation of this paper is to find a class of λ -perfect 3-graphs to support Conjecture 1. We simplify S t 3 to S t , i.e., S t = { u v i w i : 1 i t and the v i s and w i s are all distinct } . Let S 2 , t = S 2 S t . In this paper, we show that S 2 , t is λ -perfect for all t 62 , proving that (ii) of Conjecture 1 holds for G = S 2 and H = S t , also supporting (i) of Conjecture 1 in some sense.
Theorem 1.
Let t 62 be an integer and G be a 3-graph. If G is S 2 , t -free, then
λ ( G ) λ ( K 2 t + 5 3 ) = ( t + 2 ) ( 2 t + 3 ) 3 ( 2 t + 5 ) 2 .
In particular, S 2 , t is λ-perfect.
The Lagrangian density problem is strongly related to the well-known Turán problem. For a given positive integer n and an r-graph F, define the Turán number of F as the maximum number of edges attained by an n-vertex F-free r-graph, and denote it as e x ( n , F ) . The Turán density of F is defined as
π ( F ) = lim n e x ( n , F ) n r ,
such a limit is known to exist. Denote the extension of a graph F as H F , which is an r-graph obtained from F by adding ( r 2 ) new vertices to each pair { v i , v j } that is not contained in any edge of F. For example, H { 123 , 456 } = { i j v i j : 1 i 3 , 4 j 6 } , where all v i j are different. A result of Sidorenko [9,10] yields that the Lagrangian density of F is equal to the Turán density of H F . Hence, we can directly obtain the following corollary by Theorem 1.
Corollary 1.
Let t 62 be an integer. Then π ( H S 2 , t ) = 2 ( t + 2 ) ( 2 t + 3 ) ( 2 t + 5 ) 2 .
We remark that the lower bound for t can be improved slightly with some more tedious discussion, we omit it here.

2. Preliminaries

Let us list some useful results of the Lagrangian function. First, we obtain the following fact directly from the definition of Lagrangian.
Fact 1. 
Let G be an r-graph. If F is a subgraph of G, then λ ( F ) λ ( G ) .
Call an r-graph Gdense if λ ( F ) < λ ( G ) for every proper subgraph F of G. Therefore, we may assume that G is dense when we estimate the Lagrangian upper bound of an F-free r-graph G. We say that Fcovers pairs if every pair of vertices is contained in some edge of F.
Fact 2 
(Frankl and Rödl [7]). If an r-graph G is dense, then G covers pairs.
Let G be an r-graph, and i , j V ( G ) . Define
L G ( j i ) = g V ( G ) { i , j } r 1 : g { j } E ( G ) and g { i } E ( G ) .
There is a useful fact concerning the symmetry property of Lagrangian function. We omit its proof here; see [12].
Fact 3. 
Let G be a dense r-graph on vertex set [ n ] and x be an optimal weighting on G. If for some i , j [ n ] , L G ( i j ) = L G ( j i ) = . Then x i = x j .
Fact 4. 
λ ( K n r ) = n r 1 n r .
ProofofFact4. 
By Fact 3, ( 1 n , 1 n , , 1 n ) is an optimum weighting of K n r , thus we have
λ ( K n r ) = λ ( K n r , ( 1 n , 1 n , , 1 n ) ) = n r 1 n r .
Motzkin and Straus [6] proved that the Lagrangian of a simple graph G equals to the Lagrangian of its complete subgraph with maximum order, which implies a simple proof of Turán’s classical theorem. Let ω ( G ) denote the clique number of G, i.e., ω ( G ) = max { s : K s 2 G } .
Theorem 1
(Motzkin and Straus, [6]). Let G be a simple graph with ω ( G ) = t ,
λ ( G ) = λ ( K t 2 ) = 1 2 1 1 t .
However, it is not easy to determine the Lagrangian of an r-graph when r 3 . The following result is useful to calculate the Lagrangian of hypergraphs.
Lemma 1
(Frankl and Rödl [7]). Let G be an r-graph on vertex set [ n ] and x be an optimum weighting on G. Then ( w ( G , x ) x i is the partial derivative of function w ( G , x ) with respect to variable x)
w ( G , x ) x i = r λ ( G )
for every i [ n ] with x i > 0 .

3. Proof of Theorem 1

For a given r-graph G and U V ( G ) , define the link graph of U in G as the hypergraph with edge set { e V ( G ) U r | U | : e U E ( G ) } , and denote as G U . When U = { i } or U = { i , j } , we simply write as G i or G i j .
Lemma 2.
Let G be a dense 3-graph on vertex set [ n ] . If λ ( G ) > λ ( K 2 t + 5 3 ) , then for each vertex i V ( G ) , there is a clique on s vertices contained in G i with s > ( 2 t + 5 ) 2 6 t + 13 .
ProofofLemma2. 
Let x be an optimum weighting on G. Then, by Lemma 1, for every vertex i G , we have
3 λ ( K 2 t + 5 3 ) < 3 λ ( G ) = w ( G , x ) x i = w ( G i , x ) .
Note that x i > 0 , which follows from G being dense. Since G i is a simple graph, by Motzkin–Straus Theorem, w ( G i , x ) = λ ( K s ) ( 1 x i ) 3 < λ ( K s ) , where s is the clique number of G i . It follows that 3 λ ( K 2 t + 5 3 ) < λ ( K s ) . Thus, by Fact 4 and Motzkin–Straus Theorem, we have
1 2 · ( 2 t + 4 ) ( 2 t + 3 ) ( 2 t + 5 ) 2 < 1 2 · s 1 s .
It yields that 1 s < 1 ( 2 t + 4 ) ( 2 t + 3 ) ( 2 t + 5 ) 2 , i.e., 1 s < 6 t + 13 ( 2 t + 5 ) 2 . Therefore, s > ( 2 t + 5 ) 2 6 t + 13 and we are done. □
Let s be a positive integer. The s-fold enlargement of the graph F is obtained from F by adding the same s new vertices to every edge of F. For example, the 3-graph { 123 , 134 , 145 , , 1 t ( t + 1 ) } is the 1-fold enlargement of { 23 , 34 , 45 , , t ( t + 1 ) } , a path on t vertices. The following proposition is a consequence of a result of Sidorenko [10].
Proposition 1
(Sidorenko [10]). Let G be a 3-graph. If λ ( G ) > λ ( K 2 t + 5 3 ) , then G contains a copy of the 1-fold enlargement of a path on 2 t + 5 vertices.
Let A , B [ n ] ( r ) be two distinct r-set. The colex ordering on [ n ] ( r ) is the ordering satisfying that A < B if max ( ( A B ) ( B A ) ) B . For instance, 236 < 146 in N ( 3 ) . Let C ( m ; r ) be the r-graph consisting of the first m sets in the colex ordering of N ( r ) . There is a famous conjecture proposed by Frankl and Füredi [39] in 1989. They conjectured that the Lagrangian of any r-graph with m edges is no more than λ ( C ( m ; r ) ) .
Note that if m = 3 + 1 2 = + 1 3 ( 1 ) , then C ( m ; 3 ) = [ + 1 ] 3 { ( + 1 ) i : i [ 1 ] } . Clearly, { , + 1 } is not covering pairs in C ( m ; 3 ) . Thus C ( m ; 3 ) is not dense by Fact 2. Hence λ ( C ( m ; 3 ) ) λ ( K 3 ) . Moreover, we have λ ( C ( m ; 3 ) ) ) λ ( K 3 ) since K 3 C ( m ; 3 ) ) . Therefore, λ ( C ( m ; 3 ) ) ) = λ ( K 3 ) . For r = 3 , Talbot [35] first showed the conjecture holds whenever 3 m 3 + 1 2 . After then, big progress has been made in [28,29,30,36,37,40].
Theorem 2
(Talbot [35]). Let m and ℓ be integers satisfying
3 m 3 + 1 2 ,
then for any 3-graph G with m edges, λ ( G ) λ ( C ( m ; 3 ) ) . Moreover, λ ( C ( m ; 3 ) ) = λ ( K l 3 ) .
Corollary 2.
Let G be a 3-graph with m edges. If λ ( G ) > λ ( K 3 ) , then m 3 + 1 2 + 1 .
ProofofCorollary2. 
For the contrary, suppose that m l 3 + l 1 2 l . Let G be the 3-graph obtained from G by adding arbitrarily s edges to G such that l 3 e ( G ) l 3 + l 1 2 l , where s 0 is an integer. By Fact 1, we have λ ( G ) λ ( G ) . Moreover, by Theorem 2, we have λ ( G ) λ ( K l 3 ) . Thus, λ ( G ) λ ( K l 3 ) , which contradicts that λ ( G ) > λ ( K l 3 ) . □
We now give two crucial lemmas.
Lemma 3.
Let t 62 and 2 t + 6 n 2 t + 9 be two positive integers. Let G be a dense n-vertex 3-graph. If G is S 2 , t -free, then λ ( G ) λ ( K 2 t + 5 3 ) .
ProofofLemma3. 
For the contrary, suppose that λ ( G ) > λ ( K 2 t + 5 3 ) . Since λ ( G ) > λ ( K 2 t + 5 3 ) , by Proposition 1, there exists a copy of the 1-fold enlargement of a path on 2 t + 5 vertices in G. Denote this 3-graph as S with V ( S ) = { a 1 , a 2 , , a 2 t + 5 } { o } and E ( S ) = { o a 1 a 2 , o a 2 a 3 , , o a 2 t + 4 a 2 t + 5 } . Clearly, { o a 1 a 2 , , o a 2 i 1 a 2 i , , o a 2 t + 3 a 2 t + 4 } forms a copy of S t + 2 in S (see Figure 1).
For 1 k 1 < k 2 < k 3 < k 4 < k 5 t + 3 , denote
U = U ( k 1 , , k 5 ) = { a 2 k 1 1 , a 2 k 2 , a 2 k 3 1 , a 2 k 4 , a 2 k 5 1 } and H = G [ U ] .
Note that V ( S ) U contains a copy of S t with edge set i [ 6 ] E i , where
E 1 = { o a 1 a 2 , o a 3 a 4 , , o a 2 k 1 3 a 2 k 1 2 } , E 2 = { o a 2 k 1 a 2 k 1 + 1 , o a 2 k 1 + 2 a 2 k 1 + 3 , , o a 2 k 2 2 a 2 k 2 1 } , E 3 = { o a 2 k 2 + 1 a 2 k 2 + 2 , o a 2 k 2 + 3 a 2 k 2 + 4 , , o a 2 k 3 3 a 2 k 3 2 } , E 4 = { o a 2 k 3 a 2 k 3 + 1 , o a 2 k 3 + 2 a 2 k 3 + 3 , , o a 2 k 4 2 a 2 k 4 1 } , E 5 = { o a 2 k 4 + 1 a 2 k 4 + 2 , o a 2 k 4 + 3 a 2 k 4 + 4 , , o a 2 k 5 3 a 2 k 5 2 } , E 6 = { o a 2 k 5 a 2 k 5 + 1 , o a 2 k 5 + 2 a 2 k 5 + 3 , , o a 2 t + 4 a 2 t + 5 } .
Since G is S 2 , t -free, H is S 2 -free. We claim that e ( H ) 4 . Suppose that e ( H ) 5 , and relabel the vertex set of H as [ 5 ] . Without loss of generality, assume that 123 H . Since H is S 2 -free, thus there is no edge of H containing { 4 , 5 } . If there is no edge of H containing { 4 } or { 5 } , then H K 4 3 , which contradicts that e ( H ) 5 . So there are at least two edges of H such that one contains 4 and the other contains 5. Without loss of generality, assume that 124 H . Similarly, there is no edge of H containing { 3 , 5 } . So, 125 H , and thus, there is no edge of H containing { 3 , 4 } . Therefore, E ( H ) { 123 , 124 , 125 } , a contradiction. Hence, e ( H ) 4 .
There are t + 3 5 such { k 1 , k 2 , k 3 , k 4 , k 5 } , and for each e H ¯ , there are at most t 2 such { k 1 , k 2 , k 3 , k 4 , k 5 } so that e { a 2 k 1 1 , a 2 k 2 , a 2 k 3 1 , a 2 k 4 , a 2 k 5 1 } . Note that H ¯ G ¯ and e ( H ¯ ) = 5 3 e ( H ) 6 . Then
e ( G ¯ ) 6 · t + 3 5 t 2 = ( t + 3 ) ( t + 2 ) ( t + 1 ) 10 .
By Corollary 2,
e ( G ¯ ) n 3 2 t + 5 3 + 2 t + 4 2 ( 2 t + 5 ) + 1 2 t + 9 3 2 t + 5 3 + 2 t + 4 2 ( 2 t + 5 ) + 1 .
Let
f ( t ) = 2 t + 9 3 2 t + 5 3 + 2 t + 4 2 ( 2 t + 5 ) + 1 ( t + 3 ) ( t + 2 ) ( t + 1 ) 10 .
Then
f ( t ) = t 3 + 54 t 2 + 419 t + 714 10 and f ( t ) = 3 t 2 + 108 t + 419 10 .
The roots of f ( t ) = 0 are 54 4173 3 and 54 + 4173 3 . Since the quadratic function f ( t ) concave, f ( t ) is increasing in [ 0 , 54 + 4173 3 ) , and decreasing in ( 54 + 4173 3 , + ) . By direct calculation, we have 54 + 4173 3 < 62 and f ( 62 ) < 0 . Hence, f ( t ) < 0 since t 62 , that is, e ( G ¯ ) < ( t + 3 ) ( t + 2 ) ( t + 1 ) 10 . This is a contradiction with (1). We complete the proof. □
Lemma 4.
Let t 62 and n 2 t + 10 be two positive integers. Let G be a dense n-vertex 3-graph. If G is S 2 , t -free, then λ ( G ) λ ( K 2 t + 5 3 ) .
ProofofLemma4. 
For the contrary, suppose that λ ( G ) > λ ( K 2 t + 5 3 ) . For each u V ( G ) , denote a maximum clique in G u as K u . Since λ ( G ) > λ ( K 2 t + 5 3 ) with t 62 , by Lemma 2,
v ( K u ) ( 2 t + 5 ) 2 6 t + 13 43 .
Furthermore, by Proposition 1, there exists a copy of the 1-fold enlargement of a path on 2 t + 5 vertices in G. Denote this 3-graph as S with V ( S ) = { a 1 , a 2 , , a 2 t + 5 } { o } and E ( S ) = { o a 1 a 2 , o a 2 a 3 , , o a 2 t + 4 a 2 t + 5 } . Clearly, { o a 1 a 2 , , o a 2 i 1 a 2 i , , o a 2 t + 3 a 2 t + 4 } forms a copy of S t + 2 in S (see Figure 1). If we delete one vertex of { a 1 , a 2 , , a 2 t + 5 } with an odd number of subscript, then we can still find a copy of S t + 2 in S. However, If we delete one vertex of { a 1 , a 2 , , a 2 t + 5 } with an even number of subscript, then we can only guarantee that there is a copy of S t + 1 in S. The situation is always ‘worse’ when the subscript of the deleted vertex is even than when it is odd.
Since n 2 t + 10 , there are at least four vertices in V ( G ) V ( S ) , we denote four vertices among them as u 1 , u 2 , u 3 , u 4 and denote U = { u 1 , u 2 , u 3 , u 4 } . Since G is dense, then G covers pairs by Fact 2. We consider G u i u j for all 1 i < j 4 , recall that G u i u j = { v V ( G ) : v u i u j G } . □
Claim 1. 
G u i u j { o , a 2 , a 4 , , a 2 t + 4 } for all 1 i < j 4 .
ProofofClaim1. 
Suppose that there is w V ( G ) { o , a 2 , a 4 , , a 2 t + 4 } such that u 1 u 2 w G . Recall that v ( K w ) 43 , so we can pick two vertices in V ( K w ) { o , u 1 , u 2 } , say v 1 , v 2 . Thus, { u 1 u 2 w , w v 1 v 2 } forms a copy of S 2 in G { o } . We will show that there exists a copy of S t in S { w , u 1 , u 2 , v 1 , v 2 } . Recall that w { a 2 , a 4 , , a 2 t + 4 } , the worst case (For the sake of brevity of the proof, we consider only the worst case and omit other cases where { v 1 , v 2 } = { a 2 i , a 2 j + 1 } or { v 1 , v 2 } = { a 2 i + 1 , a 2 j + 1 } ) is w = a 2 k + 1 and { v 1 , v 2 } = { a 2 i , a 2 j } , where 0 k t + 2 and 1 i < j t + 2 . Assume that i < j k , delete a 2 i , a 2 j , a 2 k + 1 { a 1 , a 2 , , a 2 t + 5 } and abandon a 2 i 1 , a 2 j 1 . We can find a copy of S t with edge set { o a 1 a 2 , , o a 2 i 3 a 2 i 2 } { o a 2 i + 1 a 2 i + 2 , , o a 2 j 3 a 2 j 2 } { o a 2 j + 1 a 2 j + 2 , , o a 2 k 1 a 2 k } { o a 2 k + 2 a 2 k + 3 , , o a 2 t + 4 a 2 t + 5 } in S. So, G contains a copy of S 2 , t , a contradiction. The proofs for k < i < j or i k < j are similar. □
Claim 2. 
If u 1 u 2 a 2 G , then for each k { 1 , 3 , , 2 t + 5 } , G u 1 a k = G u 2 a k = { o } .
ProofofClaim2. 
Fix k { 1 , 3 , , 2 t + 5 } . First we prove that G u 1 a k { o , a 2 } . Suppose that there is w { o , a 2 } such that u 1 a k w G , then { a 2 u 2 u 1 , u 1 a k w } forms a copy of S 2 . It is not hard to see that there exists a copy of S t in S V ( S 2 ) . Thus, G contains a copy of S 2 , t , a contradiction. Similarly, G u 2 a k { o , a 2 } . Now we suppose that u 1 a k a 2 G . Then for k { 1 , 3 , 5 , . . . , 2 t + 5 } and k k , G u 2 a k = { o } . Otherwise { u 2 a k a 2 , a 2 a k u 1 } forms a copy of S 2 , and clearly there exists a copy of S t in S V ( S 2 ) . Thus, G contains a copy of S 2 , t , a contradiction. Let v , w V ( K a k ) { o , u 1 , u 2 , a 2 } . { u 1 a 2 a k , a k v w } forms a copy of S 2 . Our goal is to find a copy of S t in S { u 2 } { a k , a 2 , v , w } , thus, we obtain a copy of S 2 , t in G and we are done. The worst case is { v , w } = { a 2 l , a 2 l } , where 1 < l < l t + 2 . Assume that 2 l < k < 2 l . In this case k 1 , we abandon a 1 , a 2 l 1 , a 2 l + 1 and find a copy of S t 1 with edge set { o a 3 a 4 , , o a 2 l 3 a 2 l 2 } { o a 2 l + 1 a 2 l + 2 , o a k 2 a k 1 } { o a k + 1 a k + 2 , , o a 2 l 2 a 2 l 1 } { o a 2 l + 2 a 2 l + 3 , , o a 2 t + 4 a 2 t + 5 } in S. This copy of S t 1 together with { o u 2 a 1 } forms a copy of S t , and we are done. The proofs for k < 2 l < 2 l or l < l < k are similar. □
Claim 3. 
If u 1 u 2 a 2 G , then u i u j a 2 G for every pair i { 1 , 2 } , j { 3 , 4 } .
ProofofClaim3. 
Without loss of generality, assume that u 1 u 3 a 2 G . By Claim 2, u i a k o G for every pair i [ 3 ] and k { 1 , 3 , , 2 t + 5 } . Let v , w V ( K a 2 ) { a 1 , u 1 , u 2 , u 3 , u 4 , o } . Then { u 1 u 2 a 2 , a 2 v w } forms a copy of S 2 in G. Our goal is to find a copy of S t in S { u 3 } { a 2 , v , w } , thus we obtain a copy of S 2 , t in G and we are done. Similar to the proof of Claim 2, the worst case is { v , w } = { a 2 l , a 2 l } , where 1 < l < l t + 2 , and there is a copy of S t 1 not containing a 1 . This copy of S t 1 together with { o u 3 a 1 } forms a copy of S t , and we are done.
Now we continue to prove the lemma. We first claim that { u 1 u 2 o , u 3 u 4 o } or { u 1 u 3 o , u 2 u 4 o } or { u 1 u 4 o , u 2 u 3 o } is contained in G. Otherwise, without loss of generality suppose that u 1 u 2 o G . By Claim 1, we have u 1 u 2 a 2 G . If for some triple i { 1 , 2 } , j { 3 , 4 } and k { 4 , 6 , , 2 t + 4 } , u i u j a k G , then { u 1 u 2 a 2 , u i u j a k } forms a copy of S 2 in G. It is not hard to see that there is a copy of S 2 , t in S { a 2 , a k } , a contradiction. So we may assume that u i u j a k G for every triple i { 1 , 2 } , j { 3 , 4 } and k { 4 , 6 , , 2 t + 4 } . Moreover, by Claim 3, u i u j a 2 G for every pair i { 1 , 2 } , j { 3 , 4 } . Therefore, we have u 1 u 3 o , u 2 u 4 o G . Hence, we have shown that the claim holds. Without loss of generality, we assume that { u 1 u 2 o , u 3 u 4 o } is contained in G. Let c 1 , c 2 , c 3 , c 4 V ( K a 1 ) { o , u 1 , u 2 , u 3 , u 4 } . Then { c 1 c 2 a 1 , a 1 c 3 c 4 } forms a copy of S 2 . Our goal is to find a copy of S t in S { u 1 , u 2 , u 3 , u 4 } { a 1 , c 1 , c 2 , c 3 , c 4 } , thus we obtain a copy of S 2 , t in G and we are done. Similar to the proof of Claim 2, the worst case is { c 1 , c 2 , c 3 , c 4 } = { a 2 i , a 2 j , a 2 k , a 2 l } , where 1 < i < j < k < l t + 2 . It is not hard to see that there exists a copy of S t 2 in S { a 1 , c 1 , c 2 , c 3 , c 4 } . This copy of S t 1 together with { u 1 u 2 o , u 3 u 4 o } forms a copy of S t . Thus we complete the proof. □
ProofofTheorem1. 
Let G be an S 2 , t -free 3-graph with vertex set [ n ] . We can assume that G is dense, otherwise we replace G by a dense subgraph G of G with λ ( G ) = λ ( G ) . So G covers pairs by Fact 2. If n 2 t + 5 , then λ ( G ) λ ( K 2 t + 5 3 ) by Fact 1. If 2 t + 6 n 2 t + 9 , then we are done by Lemma 3. Now assume that n 2 t + 10 . For the contrary, we suppose that λ ( G ) > λ ( K 2 t + 5 3 ) . By Lemma 4, there exists a copy of S 2 , t in G, a contradiction. So π λ ( S 2 , t ) 3 ! λ ( K 2 t + 5 3 ) . Since K 2 t + 5 3 contains no S 2 , t , we have π λ ( S 2 , t ) 3 ! λ ( K 2 t + 5 3 ) . Hence, π λ ( S 2 , t ) = 3 ! λ ( K 2 t + 5 3 ) . Thus, we complete the proof. □

Author Contributions

Methodology, S.H. and B.W.; validation, S.H. and B.W.; formal analysis, S.H. and B.W.; investigation, S.H. and B.W.; resources, S.H. and B.W.; data curation, S.H. and B.W.; writing—original draft preparation, S.H. and B.W.; writing—review and editing, S.H. and B.W.; visualization, S.H. and B.W.; supervision, S.H. and B.W.; project administration, S.H. and B.W.; funding acquisition, S.H. and B.W. All authors have read and agreed to the published version of the manuscript.

Funding

The first author was partially supported by National Natural Science Foundation of China (Nos. 11931002) and China Postdoctoral Science Foundation (No. 2021M701162). The second author was partially supported by National Natural Science Foundation of China (No. 11901193) and National Natural Science Foundation of Hunan Province, China (No. 2019JJ50364).

Data Availability Statement

Not applicable.

Acknowledgments

We would like to thank Yuejian Peng for offering many constructive suggestions to our study.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Füredi, Z. Turán type problems. In Surveys in Combinatorics; London Mathematical Society Lecture Note Series 166; Cambridge University Press: Cambridge, UK, 1991; pp. 253–300. [Google Scholar]
  2. Keevash, P. Hypergrah Turán problems. In Surveys in Combinatorics; Cambridge University Press: Cambridge, UK, 2011; pp. 83–140. [Google Scholar]
  3. Bhatti, M.; Marin, M.; Zeeshan, A.; Abdelsalam, S. Recent Trends in Computational Fluid Dynamics. Front. Phys. 2020, 8, 593111. [Google Scholar] [CrossRef]
  4. Hobiny, A.; Alzahrani, F.; Abbas, I.; Marin, M. The Effect of Fractional Time Derivative of Bioheat Model in Skin Tissue Induced to Laser Irradiation. Symmetry 2020, 12, 602. [Google Scholar] [CrossRef]
  5. Zykov, A. On some properties of linear complexes. Mat. Sbornik. Novaya Seriya 1949, 24, 163. [Google Scholar]
  6. Motzkin, T.; Straus, E. Maxima for graphs and a new proof of a theorem of Turán. Can. J. Math. 1965, 17, 533–540. [Google Scholar] [CrossRef] [Green Version]
  7. Frankl, P.; Rödl, V. Hypergraphs do not jump. Combinatorica 1984, 4, 149–159. [Google Scholar] [CrossRef]
  8. Frankl, P.; Füredi, Z. Extremal problems and the Lagrange function for hypergraphs. Bull. Inst. Math. Acad. Sin. 1988, 16, 305–313. [Google Scholar]
  9. Sidorenko, A. On the maximal number of edges in a homogeneous hypergraph that does not contain prohibited subgraphs. Mat. Zametki 1987, 41, 433–455. [Google Scholar]
  10. Sidorenko, A. Asymptotic solution for a new class of forbidden r-graphs. Combinatorica 1989, 9, 207–215. [Google Scholar] [CrossRef]
  11. Hefetz, D.; Keevash, P. A hypergraph Turán theorem via Lagrangians of intersecting families. J. Comb. Theory Ser. A 2013, 120, 2020–2038. [Google Scholar] [CrossRef]
  12. Jiang, T.; Peng, Y.; Wu, B. Lagrangian densities of some sparse hypergraphs and Turań numbers of their extensions. Eur. J. Comb. 2018, 73, 20–36. [Google Scholar] [CrossRef]
  13. Bene Watts, A.; Norin, S.; Yepremyan, L. A Turán Theorem for Extensions Via an Erdos-Ko-Rado Theorem for Lagrangians. Combinatorica 2019, 39, 1149–1171. [Google Scholar] [CrossRef]
  14. Chen, P.; Liang, J.; Peng, Y. The Lagrangian density of {123,234,456} and the Turán number of its extention. Discuss. Math. Graph Theory 2021, 41, 905–921. [Google Scholar] [CrossRef]
  15. Hu, S.; Peng, Y. Lagrangian densities of enlargements of matchings in hypergraphs. Appl. Math. Comput. 2020, 374, 125068. [Google Scholar] [CrossRef]
  16. Hu, S.; Peng, Y.; Wu, B. Lagrangian densities of linear forests and Turán numbers of their extensions. J. Comb. Des. 2020, 28, 207–223. [Google Scholar] [CrossRef]
  17. Jenssen, M. Continuous Optimisation in Extremal Combinatorics. Ph.D. Thesis, London School of Economics and Political Science, London, UK, 2017. [Google Scholar]
  18. Liu, S.; Peng, Y. Generating non-jumping numbers of hypergraphs. Bull. Korean Math. Soc. 2019, 56, 1027–1039. [Google Scholar]
  19. Norin, S.; Yepremyan, L. Turán numbers of extensions. J. Comb. Theory Ser. A 2018, 155, 476–492. [Google Scholar] [CrossRef]
  20. Pikhurko, O. An exact Turán result for the generalized triangle. Combinatorica 2008, 28, 187–208. [Google Scholar] [CrossRef]
  21. Wu, B.; Peng, Y. Lagrangian densities of short 3-uniform linear paths and Turán number of their extensions. Graphs Comb. 2021, 37, 711–729. [Google Scholar] [CrossRef]
  22. Yan, Z.; Peng, Y. λ-perfect hypergraphs and Lagrangian densities of hypergraph cycles. Discret. Math. 2019, 342, 2048–2059. [Google Scholar] [CrossRef]
  23. Yan, Z.; Peng, Y. An irrational Lagrangian density of a single hypergraph. SIAM J. Discret. Math. 2022, 36, 786–822. [Google Scholar] [CrossRef]
  24. Wu, B. An Irrational Turán Density via Hypergraph Lagrangian Densities. Submitted.
  25. Baber, R.; Talbot, J. New Turán densities for 3-graphs. Electron. J. Comb. 2012, 19, 22. [Google Scholar] [CrossRef] [Green Version]
  26. Brandt, A.; Irwin, D.; Jiang, T. Stability and Turán numbers of a class of hypergraphs via Lagrangians. Comb. Probab. Comput. 2017, 26, 367–405. [Google Scholar] [CrossRef]
  27. Chen, P.; Wu, B.; Zhang, Q. A note on a conjecture of Bene Watts–Norin–Yepremyan for Lagrangian. Appl. Math. Comput. 2022, 427, 127151. [Google Scholar] [CrossRef]
  28. Gruslys, V.; Letzter, S.; Morrison, N. Hypergraph Lagrangians I: The Frankl-Füredi conjecture is false. Adv. Math. 2020, 365, 107063. [Google Scholar] [CrossRef] [Green Version]
  29. Gruslys, V.; Letzter, S.; Morrison, N. Hypergraph Lagrangians II: When colex is best. Isr. J. Math. 2021, 242, 637–662. [Google Scholar] [CrossRef]
  30. Lei, H.; Lu, L.; Peng, Y. On Lagrangians of 3-uniform hypergraphs. arXiv 2018, arXiv:1806.10846v1. [Google Scholar]
  31. Lu, L.; Wang, Z. On Hamiltonian Berge cycles in [3]-uniform hypergraphs. Discret. Math. 2021, 344, 9. [Google Scholar] [CrossRef]
  32. Mubayi, D. A hypergraph extension of Turán’s theorem. J. Comb. Theory Ser. B 2006, 96, 122–134. [Google Scholar] [CrossRef] [Green Version]
  33. Norin, S.; Yepremyan, L. Turán numbers of generalized triangles. J. Comb. Theory Ser. A 2017, 146, 312–343. [Google Scholar] [CrossRef]
  34. Gu, R.; Li, X.; Peng, Y.; Shi, Y. Some Motzkin-Straus type results for non-uniform hypergraphs. J. Comb. Optim. 2016, 31, 223–238. [Google Scholar] [CrossRef]
  35. Talbot, J. Lagrangians of hypergraphs. Comb. Probab. Comput. 2002, 11, 199–216. [Google Scholar] [CrossRef] [Green Version]
  36. Tang, Q.; Peng, Y.; Zhang, X.; Zhao, C. Connection between the clique number and the Lagrangian of 3-uniform hypergraphs. Optim. Lett. 2016, 10, 685–697. [Google Scholar] [CrossRef] [Green Version]
  37. Tyomkyn, M. Lagrangians of hypergraphs: The Frankl-Füredi conjecture holds almost everywhere. J. Lond. Math. Soc. 2017, 96, 584–600. [Google Scholar] [CrossRef] [Green Version]
  38. Wu, B.; Peng, Y. The Maximum Lagrangian of 5-uniform Hypergraphs without Containing Two Edges Intersecting at a Vertex. Acta Math. Sin. Engl. Ser. 2022, 38, 877–889. [Google Scholar] [CrossRef]
  39. Frankl, P.; Füredi, Z. Extremal problems whose solutions are the blow-ups of the small Witt-designs. J. Comb. Theory Ser. A 1989, 52, 129–147. [Google Scholar] [CrossRef]
  40. Peng, Y.; Zhao, C. A Motzkin-Straus type result for 3-uniform hypergraphs. Graphs Comb. 2013, 29, 681–694. [Google Scholar] [CrossRef]
Figure 1. A copy of S t + 2 in S.
Figure 1. A copy of S t + 2 in S.
Symmetry 14 01402 g001
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Hu, S.; Wu, B. A Note on the Lagrangian of Linear 3-Uniform Hypergraphs. Symmetry 2022, 14, 1402. https://0-doi-org.brum.beds.ac.uk/10.3390/sym14071402

AMA Style

Hu S, Wu B. A Note on the Lagrangian of Linear 3-Uniform Hypergraphs. Symmetry. 2022; 14(7):1402. https://0-doi-org.brum.beds.ac.uk/10.3390/sym14071402

Chicago/Turabian Style

Hu, Sinan, and Biao Wu. 2022. "A Note on the Lagrangian of Linear 3-Uniform Hypergraphs" Symmetry 14, no. 7: 1402. https://0-doi-org.brum.beds.ac.uk/10.3390/sym14071402

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop