arXiv is now an independent nonprofit! Learn more
License: arXiv.org perpetual non-exclusive license
arXiv:2608.24097v1 [math.CO] 25 Aug 2026

An Ore-type condition for regular factors

Jingchao Lai, Weigen Yan*†† *Corresponding Author.
   Supported by NSFC Grant (Nos. 12571366, 12071180) .
   Email address: jclai1999@163.com (J. Lai), weigenyan@263.net (W. Yan).

School of Science, Jimei University, Xiamen 361021, China

Abstract.

Let GG be a simple graph of order nn satisfying the following Ore-type condition: For any two nonadjacent vertices xx and yy of GG, dG​(x)+dG​(y)≥n+k−2d_{G}(x)+d_{G}(y)\geq n+k-2, where 1≤k≤n−11\leq k\leq n-1, k​nkn is even and dG​(x)d_{G}(x) is the degree of xx in GG. It is well known that GG has a kk-factor for k=1k=1 or 22. Lu and Ning (J. Graph Theory, 94(2020), 307-319) proved that if k≥n/2k\geq n/2, then GG has a kk-factor. In this paper, we show that GG has a kk-factor for any 1≤k≤n−11\leq k\leq n-1.

Keywords: Regular factor; Ore-type condition; Tutte’s theorem.
MSC(2020): 05C07; 05C70

1. Introduction

The graphs considered in this paper are simple if not specified. Let GG be a graph. Denote by V⁡(G)V(G), E⁡(G)E(G) and δ⁡(G)\delta(G) the vertex set, the edge set and the minimum degree of GG, respectively. Denote by |V⁡(G)||V(G)| and |E⁡(G)||E(G)| the order and size of GG, respectively. For a vertex v∈V⁡(G)v\in V(G) and a subgraph HH of GG, denote by NG​(v)N_{G}(v) the set of neighbors of vv in GG, and dG​(v)=|NG​(v)|d_{G}(v)=|N_{G}(v)|, and define NH​(v)=NG​(v)∩V⁡(H)N_{H}(v)=N_{G}(v)\cap V(H) and dH​(v)=|NH​(v)|d_{H}(v)=|N_{H}(v)|. For two disjoint subsets S,TS,T of V⁡(G)V(G), let EG​(S,T)E_{G}(S,T) denote the set of edges between SS and TT in GG, and set eG​(S,T)=|EG​(S,T)|e_{G}(S,T)=|E_{G}(S,T)|. If S={v}S=\{v\}, we write EG​(v,T)E_{G}(v,T) and eG​(v,T)e_{G}(v,T) in place of EG​({v},T)E_{G}(\{v\},T) and eG​({v},T)e_{G}(\{v\},T), respectively. For a positive integer nn, set [n]={1,2,…,n}[n]=\{1,2,\ldots,n\}. Other terminology and notation not defined here can be found in [1].

The existence of prescribed subgraphs is a central and long-standing theme in graph theory. Perhaps the best-known result is Dirac’s theorem [2] in 1952, which states that every graph of order nn contains a Hamilton cycle if δ⁡(G)≥n/2\delta(G)\geq n/2. Ore [9] later generalized Dirac’s theorem by imposing a degree sum lower bound on all pairs of nonadjacent vertices in a graph.

Let k≥1k\geq 1 be an integer. A kk-factor of a graph GG is a spanning kk-regular subgraph of GG. The seminal work of Tutte in 1952 established a complete necessary and sufficient condition for the existence of kk-factors, laying the theoretical foundation for the entire field. However, Tutte’s criterion requires verifying a combinatorial inequality over all pairs of disjoint vertex subsets, which is computationally prohibitive for large graphs. This inherent limitation has motivated decades of research devoted to deriving simple, easily verifiable sufficient conditions for the existence of kk-factors. Among all categories of sufficient conditions, degree-based criteria constitute the most extensively studied and fruitful class. Following the classical Dirac-type minimum degree paradigm, the Ore-type framework - which imposes a lower bound on the degree sum of every pair of nonadjacent vertices - also serves as a highly flexible and powerful research tool.

For a graph GG, the invariant σ2​(G)\sigma_{2}(G) is defined to be the minimum degree sum of two nonadjacent vertices of GG, i.e.,

σ2(G)=min{dG(x)+dG(y):x,y∈V(G),x≠y,xy∉E(G)}\sigma_{2}(G)=\mathrm{min}\{d_{G}(x)+d_{G}(y):x,y\in V(G),x\neq y,xy\notin E(G)\}

if GG is not complete; otherwise, let σ2​(G)=+∞\sigma_{2}(G)=+\infty.

A classical result of Ore [9] says that a simple graph GG of order nn has a Hamilton path (hence GG has a 1-factor when nn is even) if σ2​(G)≥n−1\sigma_{2}(G)\geq n-1 and GG has a Hamilton cycle (i.e., a 22-factor) if σ2​(G)≥n\sigma_{2}(G)\geq n. A natural problem is: Suppose that a simple graph GG of order nn satisfies the following Ore-type condition: σ2​(G)≥n+k−2\sigma_{2}(G)\geq n+k-2, where 1≤k≤n−11\leq k\leq n-1 and k​nkn is even. Has GG a kk-factor?

Several progressively refined degree-based results concerning the existence of kk-factors are well-known, as stated below.

Theorem 1.1.

Let k≥2k\geq 2 be an integer, and GG be a connected graph of order nn such that δ⁡(G)≥k\delta(G)\geq k and k​nkn is even. If one of the following three conditions holds, then GG has a kk-factor.
(1). δ⁡(G)≥n/2\delta(G)\geq n/2 and n≥4​k−5n\geq 4k-5. (Egawa and Enomoto [3]; Katerinis [5])
(2). σ2​(G)≥n\sigma_{2}(G)\geq n and n≥4​k−5n\geq 4k-5. (Iida and Nishimura [4])
(3). max⁡{dG​(x),dG​(y)}≥n/2\max\{d_{G}(x),d_{G}(y)\}\geq n/2 for every pair of nonadjacent vertices xx and yy of GG and n≥4​k−3n\geq 4k-3. (Nishimura [8]).

By the theorem above, if GG satisfies the Ore-type condition above, then GG has a kk-factor if k≤(n+5)/4k\leq(n+5)/4. On the other hand, Lu and Ning [7] obtained the following result for the existence of large kk-factors.

Theorem 1.2 ([7]).

Let nn and kk be two integers such that 1≤k≤n−11\leq k\leq n-1 and k​nkn is even. Let GG be a graph of order nn. If σ2​(G)≥n+k−2\sigma_{2}(G)\geq n+k-2 and k≥n/2k\geq n/2, then GG contains a kk-factor.

Hence Theorems 1.1 and 1.2 partially answer the problem above. In this paper, we obtain the following result which completely solves this problem.

Theorem 1.3.

Let nn and kk be two integers such that 1≤k≤n−11\leq k\leq n-1 and k​nkn is even. Let GG be a graph of order nn. If σ2​(G)≥n+k−2\sigma_{2}(G)\geq n+k-2, then GG contains a kk-factor.

The assumption that σ2​(G)≥n+k−2\sigma_{2}(G)\geq n+k-2 in Theorem 1.3 cannot be weakened any further. Let G=Kk−1+(Kn−k∪K1)G=K_{k-1}+(K_{n-k}\cup K_{1}), where the plus sign denotes join. Obviously, GG contains no kk-factor. However, it is easy to check that σ2​(G)=n+k−3\sigma_{2}(G)=n+k-3.

The main tools in the work of Lu and Ning [7] are Tutte’s kk-factor theorem and the Karush-Kuhn-Tucker theorem. Our main technique is to optimize and simplify the method of Lu and Ning. The rest of this paper is organized as follows. In Section 2, we introduce necessary preliminaries. In Section 3, we complete the proof of Theorem 1.3.

2. Preliminaries

In this section, we first introduce some notation and terminology related to the Tutte’s kk-factor theorem. For a graph GG and any pair of disjoint subsets S,T⊂V⁡(G)S,T\subset V(G), a component CC of G−S−TG-S-T is called a kk-odd-component if

eG​(V⁡(C),T)+k​|V⁡(C)|≡1(mod2).e_{G}(V(C),T)+k\,|V(C)|\equiv 1\pmod{2}.

Let q⁡(S,T)q(S,T) denote the number of kk-odd components of G−S−TG-S-T. Unless otherwise specified, we always let C1,…,CqC_{1},\ldots,C_{q} be all kk-odd-components of G−S−TG-S-T, and let U=⋃i=1qCiU=\bigcup_{i=1}^{q}C_{i}, where q=q⁡(S,T)q=q(S,T).

The well-known Tutte’s kk-factor theorem is stated as follows.

Theorem 2.1 (Tutte [10]).

Let kk be a positive integer. A graph GG contains no kk-factor if and only if there exist disjoint subsets S,T⊂V⁡(G)S,T\subset V(G), such that

k​|S|−k​|T|+∑x∈TdG−S​(x)−q⁡(S,T)≤−2.k\,|S|-k\,|T|+\sum_{x\in T}d_{G-S}(x)-q(S,T)\leq-2. (2.1)

From Tutte’s theorem, Katerinis and Woodall [6] deduced the following.

Theorem 2.2 (Katerinis and Woodall [6]).

Let k≥1k\geq 1 be an integer. If a graph GG contains no kk-factor, then there exist two disjoint subsets S,T⊂V⁡(G)S,T\subset V(G) such that there holds (2.1), and for all v∈V⁡(U)v\in V(U)

eG​(v,T)≤k−1,ande_{G}(v,T)\leq k-1,\quad\text{and} (2.2)
dG−S​(v)≥k+1,andd_{G-S}(v)\geq k+1,\quad\text{and} (2.3)
|V⁡(Ci)|≥3for every i∈[q].|V(C_{i})|\geq 3\quad\text{for every $i\in[q]$.} (2.4)

3. Proof of Theorem 1.3

Proof of Theorem 1.3 To prove the theorem, we argue by contradiction. Suppose that GG contains no kk-factor. By Theorem 2.2, we can choose disjoint S,T⊂V⁡(G)S,T\subset V(G) satisfying Ineqs. (2.1)-(2.4). Let C1,…,CqC_{1},\dots,C_{q} be all kk-odd components of G−S−TG-S-T, and let U=⋃i=1qCiU=\bigcup_{i=1}^{q}C_{i}. Let s=|S|s=|S|, t=|T|t=|T| and ci=|V⁡(Ci)|c_{i}=|V(C_{i})| for i∈[q]i\in[q]. Hence, by Ineqs. (2.2)-(2.4), eG​(v,T)≤k−1e_{G}(v,T)\leq k-1 and dG−S​(v)≥k+1d_{G-S}(v)\geq k+1 for every v∈V⁡(U)v\in V(U), and ci≥3c_{i}\geq 3 for every i∈[q]i\in[q]. So, |V⁡(U)|=∑i=1qci≥3​q|V(U)|=\sum_{i=1}^{q}c_{i}\geq 3q and n≥s+t+3​qn\geq s+t+3q.

If T≠∅T\neq\emptyset, then let h1:=min⁡{dG−S​(x):x∈T}h_{1}:=\min\{d_{G-S}(x):x\in T\}, and let u1∈Tu_{1}\in T such that dG−S​(u1)=h1d_{G-S}(u_{1})=h_{1}. Set P=(NG​(u1)∩T)∪{u1}P=(N_{G}(u_{1})\cap T)\cup\{u_{1}\} and p=|P|p=|P|. If T\P≠∅T\backslash P\neq\emptyset, then let h2:=min⁡{dG−S​(x):x∈T\P}h_{2}:=\min\{d_{G-S}(x):x\in T\backslash P\} and choose u2∈T\Pu_{2}\in T\backslash P such that dG−S​(u2)=h2d_{G-S}(u_{2})=h_{2}. Obviously, p=dT​(u1)+1p=d_{T}(u_{1})+1, u1​u2∉E⁡(G)u_{1}u_{2}\notin E(G) and h1≤h2h_{1}\leq h_{2}.

Claim 1.

GG is kk-connected; hence, δ⁡(G)≥k\delta(G)\geq k.

Proof.

If there exists a vertex subset WW such that G−WG-W is disconnected, then let C1′,C2′C_{1}^{\prime},C_{2}^{\prime} be two components of G−WG-W. Let x∈V⁡(C1′)x\in V(C_{1}^{\prime}) and y∈V⁡(C2′)y\in V(C_{2}^{\prime}). Then we can see that x​y∉E⁡(G)xy\notin E(G). Hence,

n+k−2≤dG​(x)+dG​(y)≤|V⁡(C1′)|−1+|W|+|V⁡(C2′)|−1+|W|.n+k-2\leq d_{G}(x)+d_{G}(y)\leq|V(C_{1}^{\prime})|-1+|W|+|V(C_{2}^{\prime})|-1+|W|.

Note that n≥|V⁡(C1′)|+|V⁡(C2′)|+|W|n\geq|V(C_{1}^{\prime})|+|V(C_{2}^{\prime})|+|W|. Hence, |W|≥k|W|\geq k, and moreover, δ⁡(G)≥k\delta(G)\geq k. ∎

Claim 2.

We have q<k​s+2q<ks+2; hence, T≠∅T\neq\emptyset and h1≤k−1h_{1}\leq k-1.

Proof.

If not, then q≥k​s+2≥2q\geq ks+2\geq 2, which implies G−S−TG-S-T is not connected. That is, S∪TS\cup T is a vertex-cut of GG. By Claim 1, we have s+t≥ks+t\geq k. Note that

∑i=1qci=|V⁡(U)|≤n−s−t≤n−k.\sum_{i=1}^{q}c_{i}=|V(U)|\leq n-s-t\leq n-k. (3.1)

Recall that |V⁡(U)|≥3​q|V(U)|\geq 3q. Then we have

n−k≥3​q≥3​k​s+6.n-k\geq 3q\geq 3ks+6. (3.2)

For every i∈[q]i\in[q], let xi∈V⁡(Ci)x_{i}\in V(C_{i}). Then NG​(xi)⊆(V⁡(Ci)∖{xi})∪S∪TN_{G}(x_{i})\subseteq(V(C_{i})\setminus\{x_{i}\})\cup S\cup T. By Ineq. (2.2),

dG​(xi)≤ci−1+s+eG​(xi,T)≤ci+s+k−2.d_{G}(x_{i})\leq c_{i}-1+s+e_{G}(x_{i},T)\leq c_{i}+s+k-2.

Hence, for any i,j∈[q]i,j\in[q] with i≠ji\neq j, we first have xi​xj∉E⁡(G)x_{i}x_{j}\notin E(G) and then

ci+s+k−2+cj+s+k−2≥dG​(xi)+dG​(xj)≥n+k−2.c_{i}+s+k-2+c_{j}+s+k-2\geq d_{G}(x_{i})+d_{G}(x_{j})\geq n+k-2.

That is, ci+cj≥n−k−2​s+2c_{i}+c_{j}\geq n-k-2s+2. Without loss of generality, we suppose that c1≤c2≤…≤cqc_{1}\leq c_{2}\leq\ldots\leq c_{q}. It is clear that

c1+c22≤1q​∑i=1qci≤n−kq.\frac{c_{1}+c_{2}}{2}\leq\frac{1}{q}\sum_{i=1}^{q}c_{i}\leq\frac{n-k}{q}.

where the second inequality holds by Ineq. (3.1). Hence we have

n−k−2​s+2≤c1+c2≤2q​(n−k).n-k-2s+2\leq c_{1}+c_{2}\leq\frac{2}{q}(n-k). (3.3)

That is, q⁡(n−k−2​s+2)≤2​(n−k)q(n-k-2s+2)\leq 2(n-k). Note that by Ineq. (3.2),

n−k−2​s+2≥3​k​s+6−2​s+2=(3​k−2)​s+8>0.n-k-2s+2\geq 3ks+6-2s+2=(3k-2)s+8>0. (3.4)

However,

q⁡(n−k−2​s+2)−2​(n−k)\displaystyle q(n-k-2s+2)-2(n-k) ≥(k​s+2)​(n−k−2​s+2)−2​(n−k)\displaystyle\geq(ks+2)(n-k-2s+2)-2(n-k) (by q≥k​s+2q\geq ks+2)
=k​s​(n−k)−(k​s+2)​(2​s−2)\displaystyle=ks(n-k)-(ks+2)(2s-2)
≥k​s​(3​k​s+6)−(k​s+2)​(2​s−2)\displaystyle\geq ks(3ks+6)-(ks+2)(2s-2) (by (3.2))
=(k​s+2)​(3​k​s−2​s+2)\displaystyle=(ks+2)(3ks-2s+2)
≥(k​s+2)​(s+2)\displaystyle\geq(ks+2)(s+2) (by k≥1k\geq 1)
>0,\displaystyle>0,

a contradiction. Thus q<k​s+2q<ks+2. It is obvious that T≠∅T\neq\emptyset. Otherwise, by Ineq. (2.1),

k​s−q≤−2,ks-q\leq-2,

a contradiction. Recalling the definition of h1h_{1}, it is clear that dG−S​(x)≥h1d_{G-S}(x)\geq h_{1} for every x∈Tx\in T. If h1≥kh_{1}\geq k, then by Ineq. (2.1),

q≥k​s−k​t+∑x∈TdG−S​(x)+2≥k​s−k​t+k​t+2=k​s+2,\displaystyle q\geq ks-kt+\sum_{x\in T}d_{G-S}(x)+2\geq ks-kt+kt+2=ks+2,

a contradiction. Hence, we have h1≤k−1h_{1}\leq k-1.

The following two claims are obvious:

Claim 3.

k≤δ⁡(G)≤dG​(u1)≤dG−S​(u1)+s=h1+sk\leq\delta(G)\leq d_{G}(u_{1})\leq d_{G-S}(u_{1})+s=h_{1}+s.

Claim 4.

p=dT​(u1)+1≤dG−S​(u1)+1=h1+1≤kp=d_{T}(u_{1})+1\leq d_{G-S}(u_{1})+1=h_{1}+1\leq k.

Finally, we require the following important result:

Claim 5.

If V⁡(U)⊈NG​(u1)V(U)\not\subseteq N_{G}(u_{1}), then

s≥k−h1+3​(q−1).s\geq k-h_{1}+3(q-1).
Proof.

Without loss of generality, take x∈V⁡(Cj)⊆V⁡(U)x\in V(C_{j})\subseteq V(U) for some j∈[q]j\in[q], and suppose that x​u1∉E⁡(G)xu_{1}\notin E(G). Recall that ci≥3c_{i}\geq 3 for every i∈[q]i\in[q]. Hence we have

cj≤n−s−t−∑i≠jci≤n−s−t−3​(q−1).c_{j}\leq n-s-t-\sum_{i\neq j}c_{i}\leq n-s-t-3(q-1). (3.5)

It is obvious that NG​(x)⊆(V⁡(Cj)∖{x})∪S∪(T∖{u1})N_{G}(x)\subseteq(V(C_{j})\setminus\{x\})\cup S\cup(T\setminus\{u_{1}\}). Furthermore, by Ineq. (3.5),

dG​(x)≤cj+s+t−2≤n−s−t−3​(q−1)+s+t−2=n−3​(q−1)−2.d_{G}(x)\leq c_{j}+s+t-2\leq n-s-t-3(q-1)+s+t-2=n-3(q-1)-2. (3.6)

By Claim 3, dG​(u1)≤h1+sd_{G}(u_{1})\leq h_{1}+s. So we have

n+k−2≤dG​(x)+dG​(u1)≤n−3​(q−1)−2+h1+s.n+k-2\leq d_{G}(x)+d_{G}(u_{1})\leq n-3(q-1)-2+h_{1}+s.

That is,

s≥k−h1+3​(q−1).s\geq k-h_{1}+3(q-1).

The result thus holds. ∎

We now consider three cases to derive a contradiction.

Case 1. T∖P=∅T\setminus P=\emptyset. That is, T=PT=P.

Let r=k−h1r=k-h_{1}. By Claims 2-4, we have r≥1r\geq 1, s≥rs\geq r and t=p≤kt=p\leq k. Furthermore, by Ineq. (2.1), we have

q\displaystyle q ≥k​s−k​t+∑x∈TdG−S​(x)+2\displaystyle\geq ks-kt+\sum_{x\in T}d_{G-S}(x)+2
≥k​s−k​t+h1​t+2\displaystyle\geq ks-kt+h_{1}t+2
=k​s−r​t+2\displaystyle=ks-rt+2 (3.7)
≥2.\displaystyle\geq 2. (by k≥tk\geq t, s≥rs\geq r)
Claim 6.

V⁡(U)⊈NG​(u1)V(U)\not\subseteq N_{G}(u_{1}).

Proof.

If V⁡(U)⊆NG​(u1)V(U)\subseteq N_{G}(u_{1}), then h1=dG−S​(u1)≥|V⁡(U)|+t−1h_{1}=d_{G-S}(u_{1})\geq|V(U)|+t-1, since T=PT=P. Note that for every x∈V⁡(U)x\in V(U), NG−S​(x)⊆(V⁡(U)∖{x})∪TN_{G-S}(x)\subseteq(V(U)\setminus\{x\})\cup T. Hence,

dG−S​(x)≤|V⁡(U)|−1+t≤h1.d_{G-S}(x)\leq|V(U)|-1+t\leq h_{1}.

But by Ineq. (2.3), dG−S​(x)≥k+1d_{G-S}(x)\geq k+1. That is, h1≥k+1h_{1}\geq k+1, which contradicts Claim 2. ∎

By Claims 6 and 5, we have s≥r+3​(q−1)s\geq r+3(q-1). Furthermore, by Ineq. (3.7),

q\displaystyle q ≥k​s−r​t+2\displaystyle\geq ks-rt+2
≥k⁡(r+3​(q−1))−r​k+2\displaystyle\geq k\big(r+3(q-1)\big)-rk+2 (by s≥r+3​(q−1),t≤ks\geq r+3(q-1),\,t\leq k)
=3​k​(q−1)+2\displaystyle=3k(q-1)+2
≥3​(q−1)+2\displaystyle\geq 3(q-1)+2 (by k≥1k\geq 1)
>q,\displaystyle>q, (by q≥2q\geq 2)

a contradiction. This completes the proof for Case 1.

Case 2. T∖P≠∅T\setminus P\neq\emptyset and h2≥kh_{2}\geq k.

Let r=k−h1r=k-h_{1}. By Claims 2-4, we have r≥1r\geq 1, s≥rs\geq r and p≤h1+1≤kp\leq h_{1}+1\leq k. Recall that h2:=min⁡{dG−S​(x):x∈T\P}h_{2}:=\min\{d_{G-S}(x):x\in T\backslash P\} if T∖P≠∅T\setminus P\neq\emptyset. By Ineq. (2.1), we have

q\displaystyle q ≥k​s−k​t+∑x∈PdG−S​(x)+∑x∈T∖PdG−S​(x)+2\displaystyle\geq ks-kt+\sum_{x\in P}d_{G-S}(x)+\sum_{x\in T\setminus P}d_{G-S}(x)+2
≥k​s−k​t+h1​p+h2​(t−p)+2\displaystyle\geq ks-kt+h_{1}p+h_{2}(t-p)+2
≥k​s−k​t+h1​p+k⁡(t−p)+2\displaystyle\geq ks-kt+h_{1}p+k(t-p)+2
=k​s−r​p+2\displaystyle=ks-rp+2 (3.8)
≥r⁡(k−p)+2\displaystyle\geq r(k-p)+2 (3.9)
≥2.\displaystyle\geq 2.
Claim 7.

V⁡(U)⊆NG​(u1)V(U)\subseteq N_{G}(u_{1}).

Proof.

If V⁡(U)⊈NG​(u1)V(U)\not\subseteq N_{G}(u_{1}), then s≥r+3​(q−1)s\geq r+3(q-1) by Claim 5. Furthermore, by Ineq. (3.8),

q\displaystyle q ≥k​s−r​p+2\displaystyle\geq ks-rp+2
≥k⁡(r+3​(q−1))−r​k+2\displaystyle\geq k\big(r+3(q-1)\big)-rk+2 (by s≥r+3​(q−1),p≤ks\geq r+3(q-1),\,p\leq k)
=3​k​(q−1)+2\displaystyle=3k(q-1)+2
≥3​(q−1)+2\displaystyle\geq 3(q-1)+2 (by k≥1k\geq 1)
>q,\displaystyle>q, (by q≥2q\geq 2)

a contradiction. The claim thus holds. ∎

Since V⁡(U)⊆NG​(u1)V(U)\subseteq N_{G}(u_{1}) and |V⁡(U)|≥3​q|V(U)|\geq 3q, we have h1=dG−S​(u1)≥3​q+p−1h_{1}=d_{G-S}(u_{1})\geq 3q+p-1. That is,

p≤h1−3​q+1.p\leq h_{1}-3q+1. (3.10)

Furthermore, by Ineqs. (3.9) and (3.10), we have

q≥r⁡(k−p)+2≥r⁡(k−h1+3​q−1)+2≥(1+3​q−1)+2>q,q\geq r(k-p)+2\geq r(k-h_{1}+3q-1)+2\geq(1+3q-1)+2>q,

where the third inequality holds since r=k−h1≥1r=k-h_{1}\geq 1, a contradiction. This completes the proof for Case 2.

Case 3. T∖P≠∅T\setminus P\neq\emptyset and h2≤k−1h_{2}\leq k-1.

Recall that u1​u2∉E⁡(G)u_{1}u_{2}\notin E(G), so we have

n+k−2≤dG​(u1)+dG​(u2)≤dG−S​(u1)+s+dG−S​(u2)+s=h1+h2+2​s.n+k-2\leq d_{G}(u_{1})+d_{G}(u_{2})\leq d_{G-S}(u_{1})+s+d_{G-S}(u_{2})+s=h_{1}+h_{2}+2s.

That is,

s≥n+k−2−h1−h22.s\geq\frac{n+k-2-h_{1}-h_{2}}{2}. (3.11)

On the other hand, by Claim 3,

s≥k−h1.s\geq k-h_{1}. (3.12)

Let λ=h22​k−h2\lambda=\frac{h_{2}}{2k-h_{2}}. Then 0≤λ<10\leq\lambda<1 since h2≤k−1h_{2}\leq k-1, and

(2​k−h2)​λ=h2​and​(2​k−h2)​(1−λ)2=k−h2.(2k-h_{2})\lambda=h_{2}\;\;\mathrm{and}\;\;\frac{(2k-h_{2})(1-\lambda)}{2}=k-h_{2}. (3.13)

By Ineqs. (3.11) and (3.12),

s=λ​s+(1−λ)​s≥λ⁡(k−h1)+(1−λ)​n+k−2−h1−h22.s=\lambda s+(1-\lambda)s\geq\lambda(k-h_{1})+(1-\lambda)\frac{n+k-2-h_{1}-h_{2}}{2}. (3.14)

Since 2​k−h2>02k-h_{2}>0, by Ineqs. (3.13) and (3.14),

(2​k−h2)​s\displaystyle(2k-h_{2})s ≥(2​k−h2)​λ​(k−h1)+(2​k−h2)​(1−λ)2​(n+k−2−h1−h2)\displaystyle\geq(2k-h_{2})\lambda(k-h_{1})+\frac{(2k-h_{2})(1-\lambda)}{2}(n+k-2-h_{1}-h_{2})
=h2​(k−h1)+(k−h2)​(n+k−2−h1−h2).\displaystyle=h_{2}(k-h_{1})+(k-h_{2})(n+k-2-h_{1}-h_{2}). (3.15)

Note that n≥s+t+|V⁡(U)|≥s+t+3​qn\geq s+t+|V(U)|\geq s+t+3q. That is,

t≤n−s−3​q.t\leq n-s-3q. (3.16)

Obviously, h1−h2≤0h_{1}-h_{2}\leq 0. By Claim 4, p≤h1+1p\leq h_{1}+1. Hence,

(h1−h2)​p≥(h1−h2)​(h1+1).(h_{1}-h_{2})p\geq(h_{1}-h_{2})(h_{1}+1). (3.17)

Furthermore, by Ineq. (2.1), we have

0\displaystyle 0 ≥k​s−k​t+∑x∈PdG−S​(x)+∑x∈T∖PdG−S​(x)+2−q\displaystyle\geq ks-kt+\sum_{x\in P}d_{G-S}(x)+\sum_{x\in T\setminus P}d_{G-S}(x)+2-q
≥k​s−k​t+h1​p+h2​(t−p)+2−q\displaystyle\geq ks-kt+h_{1}p+h_{2}(t-p)+2-q
=k​s−(k−h2)​t+(h1−h2)​p+2−q\displaystyle=ks-(k-h_{2})t+(h_{1}-h_{2})p+2-q
≥k​s−(k−h2)​(n−s−3​q)+(h1−h2)​p+2−q\displaystyle\geq ks-(k-h_{2})(n-s-3q)+(h_{1}-h_{2})p+2-q (by (3.16))
≥k​s−(k−h2)​(n−s−3​q)+(h1−h2)​(h1+1)+2−q\displaystyle\geq ks-(k-h_{2})(n-s-3q)+(h_{1}-h_{2})(h_{1}+1)+2-q (by (3.17))
=(2​k−h2)​s−(k−h2)​n+(3​(k−h2)−1)​q+(h1−h2)​(h1+1)+2\displaystyle=(2k-h_{2})s-(k-h_{2})n+\big(3(k-h_{2})-1\big)q+(h_{1}-h_{2})(h_{1}+1)+2
≥(2​k−h2)​s−(k−h2)​n+(h1−h2)​(h1+1)+2.\displaystyle\geq(2k-h_{2})s-(k-h_{2})n+(h_{1}-h_{2})(h_{1}+1)+2.

∎

Let f=(2​k−h2)​s−(k−h2)​n+(h1−h2)​(h1+1)+2f=(2k-h_{2})s-(k-h_{2})n+(h_{1}-h_{2})(h_{1}+1)+2. Then f≤0f\leq 0. However,

f\displaystyle f =(2​k−h2)​s−(k−h2)​n+(h1−h2)​(h1+1)+2\displaystyle=(2k-h_{2})s-(k-h_{2})n+(h_{1}-h_{2})(h_{1}+1)+2
≥h2​(k−h1)+(k−h2)​(n+k−2−h1−h2)−(k−h2)​n+(h1−h2)​(h1+1)+2\displaystyle\geq h_{2}(k-h_{1})+(k-h_{2})(n+k-2-h_{1}-h_{2})-(k-h_{2})n+(h_{1}-h_{2})(h_{1}+1)+2 (by (3.15))
=h2​(k−h1)+(k−h2)​(k−2−h1−h2)+(h1−h2)​(h1+1)+2.\displaystyle=h_{2}(k-h_{1})+(k-h_{2})(k-2-h_{1}-h_{2})+(h_{1}-h_{2})(h_{1}+1)+2.

Set a=k−1−h1a=k-1-h_{1} and b=k−1−h2b=k-1-h_{2}. Then a≥b≥0a\geq b\geq 0. By elementary calculations,

f\displaystyle f ≥h2​(k−h1)+(k−h2)​(k−2−h1−h2)+(h1−h2)​(h1+1)+2\displaystyle\geq h_{2}(k-h_{1})+(k-h_{2})(k-2-h_{1}-h_{2})+(h_{1}-h_{2})(h_{1}+1)+2
=1+a2−a​b+b2\displaystyle=1+a^{2}-ab+b^{2}
=1+(a−b)2+a​b\displaystyle=1+(a-b)^{2}+ab
≥1,\displaystyle\geq 1,

which contradicts f≤0f\leq 0. This completes the proof for Case 3, and hence the theorem follows.

Declaration of competing interest

There is no conflict of interest.

Data availability

No data was used for the research described in the article.

References

  • [1] J. A. Bondy, U. S. R. Murty, Graph theory, Graduate Texts in Mathematics, Springer, New York, 2008.
  • [2] G. A. Dirac, Some theorems on abstract graphs, Proc. London. Math. Soc., 2 (1952), 69-81.
  • [3] Y. Egawa, H. Enomoto, Sufficient conditions for the existence of kk-factors, Recent Studies in Graph Theory, V. R. Kulli, Ed., Vishwa International Publication, India (1989) 96-105.
  • [4] T. Iida, T. Nishimura, An Ore-type condition for the existence of kk-factors in graphs, Graphs and Combinat., 7 (1991), 353-361.
  • [5] P. Katerinis, Minimum degree of a graph and the existence of kk-factors, Proc. Indian Acad. Sci (Math. Sci.), 94 (1985), 123-127.
  • [6] P. Katerinis, D. R. Woodall, Binding numbers of graphs and the existence of kk-factors, Quart. J. Math., 38 (1987), 221-228.
  • [7] H. L. Lu, B. Ning, An Ore-type condition for large kk-factor and disjoint perfect matchings, J. Graph Theory, 94 (2020), 307-319.
  • [8] T. Nishimura, A degree condition for the existence of kk-factors, J. Graph Theory, 16 (1992), 141-151.
  • [9] O. Ore, Note on Hamilton circuits, Am. Math. Mon., 67 (1960), 55.
  • [10] W. T. Tutte, The factors of graphs, Canad. J. Math., 4 (1952), 314-328.