This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

On the sizes of bipartite 1-planar graphs thanks: This work is supported by MOE-LCSM, School of Mathematics and Statistics, Hunan Normal University, Changsha, Hunan 410081, P. R. China

Yuanqiu Huang
Department of Mathematics, Normal University of Hunan, Changsha 410081, P.R.China
hyqq@hunnu.edu.cn
Zhangdong Ouyang
Department of Mathematics, Hunan First Normal University , Changsha 410205, P.R.China
oymath@163.com
Fengming Dong
National Institute of Education, Nayang Technological University, Singapore
donggraph @163.com
Corresponding author
Abstract

A graph is called 11-planar if it admits a drawing in the plane such that each edge is crossed at most once. Let GG be a bipartite 1-planar graph with nn (4\geq 4) vertices and mm edges. Karpov showed that m3n8m\leq 3n-8 holds for even n8n\geq 8 and m3n9m\leq 3n-9 holds for odd n7n\geq 7. Czap, Przybylo and S̆krabuláková proved that if the partite sets of GG are of sizes xx and yy, then m2n+6x12m\leq 2n+6x-12 holds for 2xy2\leq x\leq y, and conjectured that m2n+4x12m\leq 2n+4x-12 holds for x3x\geq 3 and y6x12y\geq 6x-12. In this paper, we settle their conjecture and our result is even under a weaker condition 2xy2\leq x\leq y.

Keywords: bipartite graph, drawing, face, 1-planar graph.

1 Introduction

A drawing of a graph G=(V,E)G=(V,E) is a mapping DD that assigns to each vertex in VV a distinct point in the plane and to each edge uvuv in EE a continuous arc connecting D(u)D(u) and D(v)D(v). We often make no distinction between a graph-theoretical object (such as a vertex, or an edge) and its drawing. All drawings considered here are such ones that no edge crosses itself, no two edges cross more than once, and no two edges incident with the same vertex cross. The crossing number of a graph GG is the smallest number of crossings in any drawing of GG.

A drawing of a graph is 1-planar if each of its edges is crossed at most once. If a graph has a 1-planar drawing, then it is 1-planar. The notion of 1-planarity was introduced in 1965 by Ringel [12], and since then many properties of 1-planar graphs have been studied (e.g. see the survey paper [9]).

It is well-known that any simple planar graph with nn (n3n\geq 3) vertices has at most 3n63n-6 edges, and a simple and bipartite graph with nn (n3n\geq 3) vertices has at most 2n42n-4 edges. Determining the maximum number of edges in 1-planar graphs with a fixed number of vertices has aroused great interest of many authors (see, for example, [3], [5],[7], [11], [14]). It is known that [3, 7, 11] any 1-planar graph with nn (3)(\geq 3) vertices has at most 4n84n-8 edges. For bipartite 1-planar graphs, an analogous result was due to Karpov [8].

Theorem 1 ([8])

Let GG be a bipartite 11-planar graph with nn vertices. Then GG has at most 3n83n-8 edges for even n6n\not=6, and at most 3n93n-9 edges for odd nn and for n=6n=6. For all n4n\geq 4, these bounds are tight.

Note that Karpov’s upper bound on the size of a bipartite 1-planar graph is in terms of its vertex number. When the sizes of partite sets in a bipartite 1-planar graph are taken into account, Czap, Przybylo and S̆krabuláková [6] obtained another upper bound for its size (i.e., Corollary 2 in [6]).

For any graph GG, let V(G)V(G) and E(G)E(G) denote its vertex set and edge set.

Theorem 2 ([6])

If GG is a bipartite 11-planar graph with partite sets of sizes xx and yy, where 2xy2\leq x\leq y, then |E(G)|2|V(G)|+6x16|E(G)|\leq 2|V(G)|+6x-16.

For each pair of integers xx and yy with x3x\geq 3 and y6x12y\geq 6x-12, the authors in [6] constructed a bipartite 1-planar graph GG with partite sets of sizes xx and yy such that |E(G)|=2|V(G)|+4x12|E(G)|=2|V(G)|+4x-12 holds. Moreover, they believed this lower bound is optimal for such graphs and thus posed the following conjecture.

Conjecture 1 ([6])

For any integers xx and yy with x3x\geq 3 and y6x12y\geq 6x-12, if GG is a bipartite 1-planar graph with partite sets of sizes xx and yy, then |E(G)|2|V(G)|+4x12|E(G)|\leq 2|V(G)|+4x-12.

In this paper we obtain the following result which proves Conjecture 1.

Theorem 3

Let GG be a bipartite 11-planar graph with partite sets of sizes xx and yy, where 2xy2\leq x\leq y. Then |E(G)|2|V(G)|+4x12|E(G)|\leq 2|V(G)|+4x-12, and the upper bound is best possible.

The result in [6, Lemma 4] shows that the upper bound for |E(G)||E(G)| in Theorem 3 is tight. Also, if x13(y+4)x\leq\frac{1}{3}(y+4), Theorem 3 provides a better upper bound for |E(G)||E(G)| than Theorem 1.

The authors in [6] mentioned a question of Sopena [13]: How many edges we have to remove from the complete bipartite graph with given sizes of the partite sets to obtain a 1-planar graph? It is not hard to see that Theorem 3 implies the follow corollary which answers the problem.

Corollary 1

Let Kx,yK_{x,y} be the complete bipartite graph with partite sets of sizes xx and yy, where 2xy2\leq x\leq y. Then at least (x2)(y6)(x-2)(y-6) edges must be removed from Kx,yK_{x,y} such that the resulting graph becomes possibly a 1-planar graph, and the lower bound on the number of removed edges is best possible.

The remainder of this paper is arranged as follows. In Section 2, we explain some terminology and notation used in this paper. In Section 3, under some restrictions, we present several structural properties on an extension of D×D^{\times} for a 1-planar drawing DD of a bipartite 1-planar graph GG, where D×D^{\times} is a plane graph introduced in Section 2. Some important lemmas for proving Theorem 3 are given in Section 4, while the proof of this theorem is completed in Section 5. Finally, we give some further problems in Section 6.

2 Terminology and notation

All graphs considered here are simple, finite and undirected, unless otherwise stated. For terminology and notation not defined here, we refer to [2]. For any graph GG and AV(G)A\subseteq V(G), let G[A]G[A] denote the subgraph of GG with vertex set AA and edge set {eE(G):e joins two vertices in A}\{e\in E(G):e\mbox{ joins two vertices in }A\}. G[A]G[A] is called the subgraph of GG induced by AA. For a proper subset AA of V(G)V(G), let GAG-A denote the subgraph G[V(G)A]G[V(G)\setminus A].

A walk in a graph is alternately a vertex-edge sequence; the walk is closed if its original vertex and terminal vertex are the same. A path (respectively, a cycle ) of a graph is a walk (respectively, a closed walk) in which all vertices are distinct; the length of a path or cycle is the number of edges contained in it. A path (respectively, a cycle) of length kk is said to be a kk-path (respectively, kk-cycle). If a cycle CC is composed of two paths P1P_{1} and P2P_{2}, we sometimes write C=P1P2C=P_{1}\cup P_{2}.

A plane graph is a planar graph together with a drawing without crossings, and at this time we say that GG is embedded in the plane. A plane graph GG partitions the plane into a number of connected regions, each of which is called a face of GG. If a face is homeomorphic to an open disc, then it is called cellular; otherwise, noncellular. Actually, a noncellular face is homeomorphic to an open disc with a few removed “holes”. For a cellular face ff, the boundary of ff can be regarded as a closed walk of GG, while for a noncellular face ff, its boundary consists of many disjoint closed walks of GG. The size of a face is the number of the edges contained in its the boundary with each repeated edge counts twice. A face of size kk is also said to be a kk-face.

It is known that a plane graph GG has no noncellular faces if and only if GG is connected. For a connected plane graph GG, the well-known Euler’s formula states that |V(G)||E(G)|+|F(G)|=2|V(G)|-|E(G)|+|F(G)|=2, where F(G)F(G) denotes the face set of GG.

A cycle CC of a plane graph GG partitions the plane into two open regions, the bounded one (i.e., the interior of CC) and the unbounded one (i.e., the exterior of CC). We denote by int(C)int(C) and ext(C)ext(C) the interior and exterior of CC, respectively, and their closures by INT(C)INT(C) and EXT(C)EXT(C). Clearly, INT(C)EXT(C)=CINT(C)\cap EXT(C)=C. A cycle CC of a plane graph GG is said to be separating if both int(C)int(C) and ext(C)ext(C) contain at least one vertex of GG.

Let DD be a 1-planar drawing of a graph GG. The associated plane graph D×D^{\times} is the plane graph that is obtained from DD by turning all crossings of DD into new vertices of degree four; these new vertices of degree four are called the crossing vertices of D×D^{\times}.

3 An extension of D×D^{\times}

Throughout this section, we always assume that the considered graph GG (possibly disconnected) is a bipartite 1-planar graph with partite sets XX and YY, where 3|X||Y|3\leq|X|\leq|Y|. Let DD be a 1-planar drawing of GG with the minimum number of crossings, and D×D^{\times} be the associated graph of DD with the crossing vertex set WW.

Note that subsets X,YX,Y and WW form a partition of V(D×)V(D^{\times}). We color the vertices in X,YX,Y and WW by the black color, white color and red color respectively. As stated in [6], D×D^{\times} can be extended to a plane graph, denoted by DW×D^{\times}_{W}, by adding edges joining black vertices as described below:

for each vertex ww in WW, it is adjacent to two black vertices in D×D^{\times}, say x1x_{1} and x2x_{2}, and we draw an edge, denoted by ewe_{w}, joining x1x_{1} and x2x_{2} which is “most near” one side of the path x1wx2x_{1}wx_{2} of D×D^{\times} such that it does not cross with any other edge, as shown in Figure 1 (b).

Refer to caption

(a)                                                      (b)

Figure 1: The extension of D×D^{\times}.

Observe that DW×D^{\times}_{W} is a plane graph with D×D^{\times} as its spanning subgraph and the edge set of DW×D^{\times}_{W} is the union of E(D×)E(D^{\times}) and {ew:wW}\{e_{w}:w\in W\}. Although D×D^{\times} is a simple graph, DW×D^{\times}_{W} might contain parallel edges (i.e., edges with the same pair of ends), as there may exist two edges in {ew:wW}\{e_{w}:w\in W\} with the same pair of ends. An example is shown in Figure 2 (c), where DD is a 11-planar drawing of K3,6K_{3,6}.

Let FDF_{D} (or simply FF) and HDH_{D} (or simply HH) denote the subgraphs DW×[WX]D^{\times}_{W}[W\cup X] and DW×[X]D^{\times}_{W}[X] respectively. Obviously, HH is a subgraph of FF and its edge set is {ew:wW}\{e_{w}:w\in W\}, while the edge set of FF is the union of E(H)E(H) and {wx1,wx2E(D×):wW&x1,x2X}\{wx_{1},wx_{2}\in E(D^{\times}):w\in W\ \&\ x_{1},x_{2}\in X\}.

Refer to caption

(a) DD                     (b) D×D^{\times}                       (c) DW×D^{\times}_{W}                     (d) FF

Figure 2: D,D×D,D^{\times}, DW×D^{\times}_{W} and FF, where DD is a 1-planar drawing of K3,6K_{3,6}

All vertices in HH are black and the edges in HH are also called black edges. Clearly, WW is an independent set in FF and each vertex in WW (i.e., a red vertex) is of degree 22 in FF. The edges in FF incident with red vertices are called red edges. Thus, each edge in FF is either black or red, as shown in Figure 2 (d).

We have the following facts on D×D^{\times}, FF and HH:

  1. (1)

    D×D^{\times}, FF and HH are simultaneously embedded in the plane;

  2. (2)

    FF and HH are obviously loopless, but they are possibly disconnected;

  3. (3)

    weww\rightarrow e_{w} is a bijection from WW to E(H)E(H), where ww is a red vertex, and thus the number of crossings of DD equals to |E(H)||E(H)|; and

  4. (4)

    ewx1wx2e_{w}\rightarrow x_{1}wx_{2} is a bijection from E(H)E(H) to the set of 22-paths in FF whose ends are black, where ww is a red vertex and x1x_{1} and x2x_{2} are the black vertices in D×D^{\times} adjacent to ww.

Moreover we have the following propositions.

Proposition 1

Let ewe_{w} be an edge of HH with ends x1x_{1} and x2x_{2} and CC be the 33-cycle of FF consisting of ewe_{w} and its corresponding 2-path P=x1wx2P=x_{1}wx_{2}, where ww is a red vertex (see Figure 3 (a)). Then int(C)int(C) contains none of black vertices, red vertices and black edges in FF; in this sense we also say that int(C)int(C) is “empty”.

Proof.  By the definition of DW×D^{\times}_{W}, the proposition follows directly from the fact that the drawing of edge ewe_{w} is most near one side of the 2-path x1wx2x_{1}wx_{2} in D×D^{\times} without crossings with edges in D×D^{\times}. \Box

Refer to caption

(a)                                             (b)

Figure 3: Some 3-cycles and 2-cycles in FF.
Proposition 2

Assume that HH has no separating 2-cycles. If CC is a 2-cycle in HH that consists of two multiple edges e1e_{1} and e2e_{2} joining two black vertices x1x_{1} and x2x_{2} (see Figure 3 (b)), then either int(C)int(C) or ext(C)ext(C) contains neither black vertices nor red vertices.

Proof.  As HH has no separating 2-cycles, either int(C)int(C) or ext(C)ext(C) contains no black vertices. Assume that int(C)int(C) does not contain black vertices.

Suppose that int(C)int(C) contains red vertices. Then, int(C)int(C) contains white vertices of D×D^{\times}. As int(C)int(C) does not contain black vertices, each white vertex in int(C)int(C) is of degree at most 22 in D×D^{\times}. Thus, we can redraw the edges of DD in int(C)int(C) such that these edges make no crossings, and then obtain a 1-planar drawing of GG with fewer crossings than DD, contradicting to the choice of DD. Hence int(C)int(C) does not contain red vertices and the conclusion holds. \Box

Proposition 3

Assume that HH contains no separating 2-cycles. Then the edge multiplicity of HH is at most 2.

Proof.  Assume to contrary that HH has three multiple edges e1e_{1}, e2e_{2} and e3e_{3} which join the same pair of black vertices x1x_{1} and x2x_{2}. Then these three edges divide the plane into three regions, denoted by α\alpha, β\beta and γ\gamma, as shown in Figure 4 (a). By Proposition 2, at least two of these three regions contain neither red vertices nor black vertices, except on its boundary. We may assume α\alpha and γ\gamma are such two regions.

Let P=x1wx2P=x_{1}wx_{2} be the 2-path of FF that corresponds to edge e3e_{3}, where ww is a red vertex. Thus, this path must be within region β\beta, as shown in Figure 4 (b).

As PP is within region β\beta, black edges e1e_{1} and e2e_{2} are in different sets int(e3P)int(e_{3}\cup P) and ext(e3P)ext(e_{3}\cup P), a contradiction to Proposition 1. The proof is then completed. \Box

Refer to caption

(a)                                                  (b)

Figure 4: Possible three multiple edges.

An edge of HH is called a simple edge if it is not parallel to another edge in HH and a partnered edge otherwise. It follows from Proposition 3 that, if HH has no separating 2-cycles, then each partnered edge ee in HH is parallel to a unique partnered edge ee^{\prime} in HH.

Let CC be a cycle and PP be a path in HH such that the end vertices of PP are the only vertices in both CC and PP. When we say that PP lies in int(C)int(C) (resp. ext(C)ext(C)), it means that all edges and internal vertices of PP lie in int(C)int(C) (resp. ext(C)ext(C)).

Proposition 4

Assume that HH has no separating 22-cycles. Let CC be a 33-cycle of HH consisting of black vertices x1x_{1}, x2x_{2} and x3x_{3}, and ee be the edge on CC joining x1x_{1} and x3x_{3}. Assume that ee^{\prime} is a partnered edge in HH which is parallel to ee. If P=x1wx3P=x_{1}wx_{3} and P=x1wx3P^{\prime}=x_{1}w^{\prime}x_{3} are the 22-paths in FF corresponding to ee and ee^{\prime} respectively, then one of PP and PP^{\prime} lies in int(C)int(C) and the other in ext(C)ext(C).

Refer to caption

(a) F[{x1,x3,w,w}]F[\{x_{1},x_{3},w,w^{\prime}\}]                            (b) F[{x1,x2,x3,w,w}]F[\{x_{1},x_{2},x_{3},w,w^{\prime}\}]

Figure 5: x2x_{2} lies in ext(C2)ext(C_{2}), where C2C_{2} is the cycle x1wx3wx1x_{1}wx_{3}w^{\prime}x_{1}

Proof.  Let C0C_{0} denote the 2-cycle of HH consisting of edges ee and ee^{\prime}. By Proposition 2, we may assume that int(C0)int(C_{0}) contains neither black vertices nor red vertices. Thus, both ww and ww^{\prime} are in ext(C0)ext(C_{0}).

Let C1C_{1} denote the 33-cycle of FF consisting of edge ee and path PP and C1C^{\prime}_{1} the 33-cycle of FF consisting of edge ee^{\prime} and path PP^{\prime}. By Proposition 1, both int(C1)int(C_{1}) and int(C1)int(C^{\prime}_{1}) are empty. Thus, the subgraph F[{x1,x3,w,w}]F[\{x_{1},x_{3},w,w^{\prime}\}] is as shown in Figure 5 (a).

As these three regions int(C0)int(C_{0}), int(C1)int(C_{1}) and int(C1)int(C^{\prime}_{1}) do not contain black vertices, x2x_{2} must be in ext(C2)ext(C_{2}), where C2C_{2} is the 44-cycle of FF consisting of paths P=x1wx3P=x_{1}wx_{3} and P=x1wx3P^{\prime}=x_{1}w^{\prime}x_{3}. As FF is a plane graph, path x1x2x3x_{1}x_{2}x_{3} must lies in ext(C2)ext(C_{2}), as shown in Figure 5 (b).

Hence the conclusion holds. \Box

Proposition 5

Suppose that HH has no separating 2-cycles. For any 33-cycle CC in HH, if int(C)int(C) contains exactly rr red vertices, where 0r20\leq r\leq 2, then CC contains at least 3r3-r simple edges of FF.

Proof.  Let e1,e2e_{1},e_{2} and e3e_{3} be the three edges on CC. Suppose that eie_{i} is not a simple edge of HH, where 1i31\leq i\leq 3. Then eie_{i} is parallel to another partnered edge eie^{\prime}_{i} of HH. Let PiP_{i} and PiP^{\prime}_{i} be 2-paths in FF which correspond to edges eie_{i} and eie^{\prime}_{i} respectively. Since HH has no separating 2-cycles, by Proposition 4, int(C)int(C) contains a red vertex that is on PiP_{i} or PiP^{\prime}_{i}.

The above conclusion implies that the number of red vertices in int(C)int(C) is not less than the number of partnered edges on CC. Thus, the result holds. \Box

4 Some lemmas

Let GG be a bipartite graph with partite sets XX and YY and 𝒪{\cal O} be a disk on the plane. If DD is a 11-planar drawing of GG that draws all vertices of XX on the boundary of 𝒪{\cal O} and all vertices of YY and all edges of GG in the interior of 𝒪{\cal O}, then we say that DD is a 11-disc 𝒪X{\cal O}_{X} drawing of GG.

Lemma 1

Let GG be a bipartite graph with partite sets XX and YY, and let DD be a 1-disc 𝒪X{\cal O}_{X} drawing of GG with the minimum number of crossings kk. If |X|=3|X|=3, then k{0,1,3}k\in\{0,1,3\} and |E(G)|2|Y|+1+k|E(G)|\leq 2|Y|+1+\big{\lceil}\sqrt{k}\big{\rceil}, i.e.,

|E(G)|{2|Y|+1,if k=0;2|Y|+2,if k=1;2|Y|+3,if k=3.|E(G)|\leq\left\{\begin{array}[]{ll}2|Y|+1,\quad\mbox{if }k=0;\\ 2|Y|+2,\quad\mbox{if }k=1;\\ 2|Y|+3,\quad\mbox{if }k=3.\end{array}\right.
Refer to caption

(a) |Y3|=2|Y_{3}|=2                            (b) |Y3|=3|Y_{3}|=3

Figure 6: The 1-disc 𝒪X{\cal O}_{X} drawing of GG for Y=Y3Y=Y_{3} and |Y3|{2,3}|Y_{3}|\in\{2,3\}

Proof.  Assume that |X|=3|X|=3. For any integer i0i\geq 0, let YiY_{i} be the set of vertices yy in YY with dG(y)=id_{G}(y)=i. As |X|=3|X|=3 and YY is independent in GG, Yi=Y_{i}=\emptyset holds for all i4i\geq 4.

It can be checked easily that, for each vertex yy in Y\in Y, if yY3y\notin Y_{3}, then yy is not incident with any crossed edge. Thus, Gi2YiG-\bigcup_{i\leq 2}Y_{i} has exactly kk crossings, and it suffices to show that |Y3|3|Y_{3}|\leq 3 and

k={0,if |Y3|1;1,if |Y3|=2;3,if |Y3|=3.k=\left\{\begin{array}[]{ll}0,&\mbox{if }|Y_{3}|\leq 1;\\ 1,&\mbox{if }|Y_{3}|=2;\\ 3,&\mbox{if }|Y_{3}|=3.\end{array}\right.

The rest of the proof will be completed by showing the following claims.

Claim (a): |Y3|3|Y_{3}|\leq 3.

Suppose that |Y3|4|Y_{3}|\geq 4. Then, there exists a bipartite 11-planar drawing DD^{\prime} isomorphic to K3,2|Y3|K_{3,2|Y_{3}|} obtained from D[XY3]D[X\cup Y_{3}] by copying all vertices and edges in the interior of 𝒪X{\cal O}_{X} to its exterior, implying that K3,8K_{3,8} is 1-planar. It is a contradiction to the fact that K3,7K_{3,7} is not 1-planar due to Czap and Hudák [4].

Thus, Claim (a) holds.

Claim (b): If |Y3|1|Y_{3}|\leq 1, then Gi2YiG-\bigcup_{i\leq 2}Y_{i} has no crossings, i.e., k=0k=0.

Claim (b) can be verified easily.

Claim (c): For any two vertices y1,y2Y3y_{1},y_{2}\in Y_{3}, some edge incident with y1y_{1} crosses with some edge incident with y2y_{2}, as shown in Figure 6 (a).

If Claim (c) fails, then G[X{y1,y2}]G[X\cup\{y_{1},y_{2}\}] is a plane graph and we can get a drawing of K3,3K_{3,3} from G[X{y1,y2}]G[X\cup\{y_{1},y_{2}\}] by adding a new vertex yy^{\prime} and three edges joining yy^{\prime} to all vertices in XX in the exterior of 𝒪X{\cal O}_{X} without any crossing, implying that K3,3K_{3,3} is planar, a contradiction.

Claim (d): k=1k=1 when |Y3|=2|Y_{3}|=2, and k=3k=3 when |Y3|=3|Y_{3}|=3.

By Claim (c), k(|Y3|2)k\geq{|Y_{3}|\choose 2}. By the drawings in Figure 6, k1k\leq 1 when |Y3|2|Y_{3}|\leq 2, and k3k\leq 3 when |Y3|3|Y_{3}|\leq 3. Thus, Claim (d) holds.

The result follows from Claims (a), (b) and (d). \Box

Lemma 2

Let GG be a plane simple graph with |V(G)|3|V(G)|\geq 3. If GG has exactly cc components and tt (0)(\geq 0) cellular 33-faces, then |E(G)|2|V(G)|3c+t2|E(G)|\leq 2|V(G)|-3-c+\frac{t}{2}.

Proof.  If c=1c=1, since GG is simple, each face of HH is a cellular face and has size at least 3. Then, in this case, the conclusion can be proved easily by applying the Euler’s formula.

Now we assume that c2c\geq 2. We can obtain a simple and connected plane graph GG^{\prime} from GG by adding c1c-1 edges.

For every noncellular face FF of GG, we assume that its boundary consists of \ell disjoint closed walks of GG, and then we can add 1\ell-1 new edges (not add the vertex) by appropriately drawing these new edges within FF so that FF is transformed into a cellular face of size at least 4 because |V(G)|3|V(G)|\geq 3. Therefore, the resulting graph GG^{\prime} is a simple and connected plane, and all faces of GG^{\prime} are cellular.

Note that adding the c1c-1 new edges does not produce new cellular 3-faces, and thus GG^{\prime} has exactly tt faces of size 3. The conclusion for connected plane graphs implies that

|E(G)|2|V(G)|4+t/2.|E(G^{\prime})|\leq 2|V(G^{\prime})|-4+t/2.

As V(G)=V(G)V(G^{\prime})=V(G) and |E(G)|=|E(G)|+c1|E(G^{\prime})|=|E(G)|+c-1, the above inequality implies that |E(G)|2|V(G)|+t/23c|E(G)|\leq 2|V(G)|+t/2-3-c. \Box

Lemma 3

Let GG be a simple and bipartite plane graph with |V(G)|3|V(G)|\geq 3. If GG has exactly cc components and tt cellular faces whose boundaries are of length at least 66, then |E(G)|2|V(G)|3ct|E(G)|\leq 2|V(G)|-3-c-t.

Proof.  If GG is connected (i.e. c=1c=1), since GG is bipartite and simple, then each face of GG is a cellular face, and has the size at least 4. Because GG has tt faces of size at least 6, it follows from the Euler’s formula that |E(G)|2|V(G)|4t|E(G)|\leq 2|V(G)|-4-t.

Now assume that c2c\geq 2. We can obtain a simple and connected bipartite plane graph GG^{\prime} from GG by adding c1c-1 edges.

For every noncellular face FF of GG consisting of \ell distinct closed walks, similar to the proof of Lemma 2, we can add 1\ell-1 new edges within this noncellular face so that FF is transformed into a cellular face. We can ensure that those new added edges join the vertices in different partite sets of GG. Hence the resulting plane graph GG^{\prime} is simple, bipartite and connected. Clearly, all faces of GG^{\prime} are cellular, and GG^{\prime} has at least tt faces whose boundaries are of length at least 66. The conclusion for bipartite and connected plane graphs implies that

|E(G)|2|V(G)|4t.|E(G^{\prime})|\leq 2|V(G^{\prime})|-4-t.

As V(G)=V(G)V(G^{\prime})=V(G) and |E(G)|=|E(G)|+c1|E(G^{\prime})|=|E(G)|+c-1, the above inequality implies that |E(G)|2|V(G)|3ct|E(G)|\leq 2|V(G)|-3-c-t. \Box

Remark: Lemmas 2 and 3 can be strengthened when GG contains isolated vertices. Let V1(G)V_{\geq 1}(G) be the set of non-isolated vertices in GG. Then, under the condition |V1(G)|3|V_{\geq 1}(G)|\geq 3, the conclusions of both Lemmas 2 and 3 still hold after |V(G)||V(G)| is replaced by |V1(G)||V_{\geq 1}(G)|.

5 Proof of Theorem 3

The whole section contributes to the proof of Theorem 3.

Proof of Theorem 3. Suppose that Theorem 3 fails and χ\chi is the minimum integer with χ2\chi\geq 2 such that for some bipartite 1-planar graph GG with partite sets XX and YY, where χ=|X||Y|\chi=|X|\leq|Y|, |E(G)|>2|V(G)|+4|X|12|E(G)|>2|V(G)|+4|X|-12 holds.

We will prove the following claims to show that this assumption leads to a contradiction.

Claim 1: χ4\chi\geq 4.

Proof. Let GG be any bipartite 11-planar graph with bipartitions XX and YY, where 2|X||Y|2\leq|X|\leq|Y|. If |X|=2|X|=2, obviously,

|E(G)|2|Y|=2(2+|Y|)+4×212=2|V(G)|+4|X|12.|E(G)|\leq 2|Y|=2(2+|Y|)+4\times 2-12=2|V(G)|+4|X|-12.

Now assume that |X|=3|X|=3. Let YiY_{i} be the set of vertices yy in YY with dG(y)=id_{G}(y)=i. Then |E(G)|2|Y|+|Y3||E(G)|\leq 2|Y|+|Y_{3}|. Since the complete bipartite graph K3,7K_{3,7} is not 1-planar (see [4]), we have |Y3|6|Y_{3}|\leq 6, implying that |E(G)|2|Y|+6|E(G)|\leq 2|Y|+6. As x=3x=3, we have

2|V(G)|+4|X|12=6|X|+2|Y|12=2|Y|+6.2|V(G)|+4|X|-12=6|X|+2|Y|-12=2|Y|+6.

Thus, |E(G)|2|V(G)|+4|X|12|E(G)|\leq 2|V(G)|+4|X|-12.

By the assumption of χ\chi, we have χ4\chi\geq 4. \Box

In the following, we assume that GG is a bipartite 11-planar graph with bipartitions XX and YY, where χ=|X||Y|\chi=|X|\leq|Y|, such that

|E(G)|>2|V(G)|+4|X|12.|E(G)|>2|V(G)|+4|X|-12. (1)

Let DD be a 1-planar drawing of GG with the minimum number of crossings and WW be the set of its crossings. Introduced in Section 3, DW×D^{\times}_{W} is a plane graph extended from D×D^{\times}, and FF and HH are the subgraphs DW×[XW]D^{\times}_{W}[X\cup W] and DW×[X]D^{\times}_{W}[X] of DW×D^{\times}_{W} respectively. All vertices in XX are black vertices, all vertices in YY are white vertices and all vertices in WW are red vertices.

We are now going to prove the following claim.

Claim 2: HH has no separating 2-cycles.

Proof. Assume to the contrary that HH has a separating 2-cycle CC consisting of two parallel edges e1e_{1} and e2e_{2} joining black vertices x1x_{1} and x2x_{2} (see Figure 3 (b)), such that both int(C)int(C) and ext(C)ext(C) contain black vertices.

Let G1=GINT(C)G_{1}=G\bigcap INT(C) and G2=GEXT(C)G_{2}=G\bigcap EXT(C). Obviously, both G1G_{1} and G2G_{2} are bipartite 1-planar subgraphs of GG. Moreover, we can see that G1G2=GG_{1}\cup G_{2}=G, and G1G2={x1,x2}G_{1}\cap G_{2}=\{x_{1},x_{2}\}.

For i=1,2i=1,2, GiG_{i} has a bipartition XiX_{i} and YiY_{i}, where Xi=XV(Gi)X_{i}=X\cap V(G_{i}) and Yi=YV(Gi)Y_{i}=Y\cap V(G_{i}). Clearly, |X1|+|X2|=|X|+2=χ+2|X_{1}|+|X_{2}|=|X|+2=\chi+2 and |Y1|+|Y2|=|Y||Y_{1}|+|Y_{2}|=|Y|.

Since CC is a separating cycle of HH, both int(C)int(C) and ext(C)ext(C) contain black vertices, implying that |Xi|3|X_{i}|\geq 3 for both i=1,2i=1,2. As |X1|+|X2|=|X|+2|X_{1}|+|X_{2}|=|X|+2, we have |Xi|<|X|=χ|X_{i}|<|X|=\chi and so min{|Xi|,|Yi|}|Xi|<χ\min\{|X_{i}|,|Y_{i}|\}\leq|X_{i}|<\chi.

For i=1,2i=1,2, if |Yi|1|Y_{i}|\leq 1, then |E(Gi)||Yi||Xi||E(G_{i})|\leq|Y_{i}|\cdot|X_{i}| and |E(Gi)|2|V(Gi)|+4|Xi|12|E(G_{i})|\leq 2|V(G_{i})|+4|X_{i}|-12 holds; if |Yi|2|Y_{i}|\geq 2, then 2min{|Xi|,|Yi|}|Xi|<χ2\leq\min\{|X_{i}|,|Y_{i}|\}\leq|X_{i}|<\chi and the assumption on χ\chi implies that the conclusion of Theorem 3 holds for GiG_{i}, i.e.,

|E(Gi)|2|V(Gi)|+4min{|Xi|,|Yi|}122|V(Gi)|+4|Xi|12.\displaystyle|E(G_{i})|\leq 2|V(G_{i})|+4\min\{|X_{i}|,|Y_{i}|\}-12\leq 2|V(G_{i})|+4|X_{i}|-12. (2)

Thus, by (2),

|E(G)|\displaystyle|E(G)| =\displaystyle= |E(G1)|+|E(G2)|\displaystyle|E(G_{1})|+|E(G_{2})| (3)
\displaystyle\leq 2(|V(G1)|+|V(G2)|)+4(|X1|+|X2|)24\displaystyle 2(|V(G_{1})|+|V(G_{2})|)+4(|X_{1}|+|X_{2}|)-24
=\displaystyle= 2(|V(G)|+2)+4(|X|+2)24\displaystyle 2(|V(G)|+2)+4(|X|+2)-24
=\displaystyle= 2|V(G)|+4|X|12,\displaystyle 2|V(G)|+4|X|-12,

which contradicts to the assumption in (1).

Hence Claim 2 holds. \Box

It is known from Proposition 3 that the edge multiplicity of each edge in HH is at most 2. Then, there exists a subset AA of E(H)E(H) such that both HAH\langle A\rangle and HAH-A are simple graphs and each edge in AA is parallel to some edge in E(H)AE(H)-A, where HAH\langle A\rangle is the spanning subgraph of HH with edge set AA and HAH-A is the graph obtained from HH by removing all edges in AA. Clearly, |A||A| is the number of pairs of edges ee and ee^{\prime} in HH which are parallel.

Let HH^{\prime} denote HAH-A. Obviously, |A||E(H)||A|\leq|E(H^{\prime})|, and

|E(H)|=|E(H)|+|A|.|E(H)|=|E(H^{\prime})|+|A|. (4)

Claim 3: HH^{\prime} contains at least one cellular 3-face.

Proof. Suppose that HH^{\prime} has no cellular 3-faces. As HH^{\prime} is a simple plane graph, by Lemma 2, |E(H)|2|V(H)|4=2|X|4|E(H^{\prime})|\leq 2|V(H^{\prime})|-4=2|X|-4. Because |A||E(H)||A|\leq|E(H^{\prime})|, it follows from (4) that |E(H)|2|E(H)|4|X|8|E(H)|\leq 2|E(H^{\prime})|\leq 4|X|-8.

Since each edge of HH is in 1-1 correspondence with a crossing of a drawing DD of GG, we can obtain a simple bipartite plane graph (possibly disconnected), denoted by GG^{\prime}, by removing |E(H)||E(H)| edges from GG each of which is a crossed edge of GG. By Lemma 3, |E(G)|2|V(G)|4=2|V(G)|4|E(G^{\prime})|\leq 2|V(G^{\prime})|-4=2|V(G)|-4. Therefore,

|E(G)|=|E(G)|+|E(H)||E(G)|+4|X|8=2|V(G)|+4|X|12,|E(G)|=|E(G^{\prime})|+|E(H)|\leq|E(G^{\prime})|+4|X|-8=2|V(G)|+4|X|-12,

a contradiction to the assumption in (1).

Hence Claim 3 holds. \Box

Now we assume that HH^{\prime} has exactly tt cellular 3-faces, where t1t\geq 1. Let 𝒯(H)\mathscr{T}(H^{\prime}) denote the set of cellular 3-faces in HH^{\prime}. So t=|𝒯(H)|t=|\mathscr{T}(H^{\prime})|.

For each Δ𝒯(H)\Delta\in\mathscr{T}(H^{\prime}), for convenience we also use “Δ\Delta” to represent the 3-cycle corresponding the boundary of Δ\Delta if there is no confusions in the context. Let GΔ=GINT(Δ)G_{\Delta}=G\bigcap INT(\Delta). Since Δ\Delta is a cellular 3-face of HH^{\prime}, there are no black vertices lying in int(Δ)int(\Delta), and thus GΔG_{\Delta} is a bipartite with with exactly three black vertices, which lie on the boundary of the face Δ\Delta.

Let Δ𝒯(H)\Delta\in\mathscr{T}(H^{\prime}). Since DD is a 1-planar drawing of GG with minimal number of crossings, the induced subdrawing of DD of GΔG_{\Delta} is a 1-disc 𝒪XΔ{\cal O}_{X_{\Delta}} drawing of GΔG_{\Delta} with the minimum number of crossings, where XΔX_{\Delta} is the set of three black vertices on the boundary of Δ\Delta. Otherwise, we redraw the edges of GG lying in the interior of Δ\Delta, and obtain a 1-planar drawing of GG with fewer crossings than DD, contradicting to the choice of DD. By Lemma 1, the number of crossings of DD in int(Δ)int(\Delta) is a number in the set {0,1,3}\{0,1,3\}.

For any j{0,1,3}j\in\{0,1,3\}, let 𝒯(j)(H)\mathscr{T}^{(j)}(H^{\prime}) be the set of members Δ\Delta in 𝒯(H)\mathscr{T}(H^{\prime}) such that int(Δ)int(\Delta) contains exactly jj crossings of DD. Assume that 𝒯(j)(H)={Δi(j):1itj}\mathscr{T}^{(j)}(H^{\prime})=\{\Delta^{(j)}_{i}:1\leq i\leq t_{j}\}, where tj=|𝒯(j)(H)|t_{j}=|\mathscr{T}^{(j)}(H^{\prime})|. Thus, t0+t1+t3=tt_{0}+t_{1}+t_{3}=t.

For each Δi(j)𝒯(j)(H)\Delta^{(j)}_{i}\in\mathscr{T}^{(j)}(H^{\prime}), let ei(j)e^{(j)}_{i} be the number of the edges of the graph GΔi(j)G_{\Delta^{(j)}_{i}} and yi(j)y^{(j)}_{i} be the number of white vertices in int(Δi(j))int(\Delta^{(j)}_{i}).

Claim 4: j{0,1,3}i=1tjei(j)2j{0,1,3}i=1tjyi(j)+(t0+2t1+3t3)\sum\limits_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}e^{(j)}_{i}\leq 2\sum\limits_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}y^{(j)}_{i}+(t_{0}+2t_{1}+3t_{3}).

Proof. By Lemma 1, ei(j)2yi(j)+1+je^{(j)}_{i}\leq 2y^{(j)}_{i}+1+\big{\lceil}\sqrt{j}\big{\rceil} holds for any j{0,1,3}j\in\{0,1,3\} and 1itj1\leq i\leq t_{j}. Thus, Claim 4 holds. \Box

Claim 5: |E(H)|4|X|8+t(3t0+2t1)/2|E(H)|\leq 4|X|-8+t-({3t_{0}}+2t_{1})/2.

Proof. Since HH^{\prime} is a simple plane graph with exactly t1t\geq 1 cellular 3-faces and |V(H)|=|X|=χ4|V(H^{\prime})|=|X|=\chi\geq 4, by Lemma 2,

|E(H)|2|V(H)|4+t2=2|X|4+t2.|E(H^{\prime})|\leq 2|V(H^{\prime})|-4+\frac{t}{2}=2|X|-4+\frac{t}{2}. (5)

On the other hand, by Proposition 5, for any j{0,1,3}j\in\{0,1,3\} and 1itj1\leq i\leq t_{j}, at least 3j3-j edges on the boundary of Δi(j)\Delta^{(j)}_{i} are simple edges, implying that at least 3j3-j edges on the boundary of Δi(j)\Delta^{(j)}_{i} are in HH^{\prime}. Because each simple edge of HH^{\prime} belongs to the boundaries of at most two different faces of HH^{\prime}, it follows that |E(H)|(3t0+2t1)/2|E(H^{\prime})|\geq(3t_{0}+2t_{1})/2. Then, by (4),

|A||E(H)|(3t0+2t1)/2,|A|\leq|E(H^{\prime})|-(3t_{0}+2t_{1})/2,

and therefore, by (4) and (5),

|E(H)|=|E(H)|+|A|4|X|8+t(3t0+2t1)/2.|E(H)|=|E(H^{\prime})|+|A|\leq 4|X|-8+t-(3t_{0}+2t_{1})/2. (6)

Thus, Claim 5 holds. \Box

Let DD^{\prime} denote the drawing obtained from DD by deleting all white vertices and edges of DD that lie in the interiors of all cellular 3-faces Δi(j)\Delta^{(j)}_{i} of HH^{\prime}, where j{0,1,3}j\in\{0,1,3\} and 1itj1\leq i\leq t_{j}, and let GG^{\prime} denote the graph represented by DD^{\prime}.

We see that the graph GG^{\prime} is a bipartite 1-planar graph with a bipartition XX and Y=YV(G)Y^{\prime}=Y\cap V(G^{\prime}), where |Y|=|Y|j{0,1,3}i=1tjyi(j)|Y^{\prime}|=|Y|-\sum\limits_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}y^{(j)}_{i}. Thus,

|V(G)|=|V(G)|j{0,1,3}i=1tjyi(j).|V(G^{\prime})|=|V(G)|-\sum_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}y^{(j)}_{i}. (7)

and

|E(G)|=|E(G)|j{0,1,3}i=1tjei(j).|E(G^{\prime})|=|E(G)|-\sum_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}e^{(j)}_{i}. (8)

As the number of crossings of DD equals to |E(H)||E(H)| and DD^{\prime} has no crossings lying in the interior of any cellular 3-face of HH^{\prime}, DD^{\prime} has exactly |E(H)|(t1+3t3)|E(H)|-(t_{1}+3t_{3}) crossings.

For each crossing of DD^{\prime}, we remove exactly one crossed edge from GG^{\prime} and obtain a bipartite plane graph GG^{*}. Thus, |E(G)|=|E(G)|(|E(H)|(t1+3t3))|E(G^{*})|=|E(G^{\prime})|-(|E(H)|-(t_{1}+3t_{3})). Then, (8) implies that

|E(G)|\displaystyle|E(G^{*})| =\displaystyle= (|E(G)|j{0,1,3}i=1tjei(j))|E(H)|+(t1+3t3).\displaystyle\Big{(}|E(G)|-\sum_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}e^{(j)}_{i}\Big{)}-|E(H)|+(t_{1}+3t_{3}). (9)

Clearly, by (7),

|V(G)|=|V(G)|+j{0,1,3}i=1tjyi(j)=|V(G)|+j{0,1,3}i=1tjyi(j).|V(G)|=|V(G^{\prime})|+\sum_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}y^{(j)}_{i}=|V(G^{*})|+\sum_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}y^{(j)}_{i}. (10)

Now, we shall obtain an upper bound of |E(G)||E(G^{*})| in terms of |V(G)||V(G^{*})| by constructing a bipartite plane graph with at least tt cellular 6-faces.

Claim 6: |E(G)|2|V(G)|4t|E(G^{*})|\leq 2|V(G^{*})|-4-t.

Proof. Note that the simple and bipartite plane graph GG^{*} is obtained from GG by removing all white vertices and edges of GG lying in the interiors of all cellular 3-faces of HH^{\prime} and, for each crossing of DD not lying in any cellular 3-face of HH^{\prime}, removing exactly one edge of GG involved in this crossing.

Now let GG^{**} denote the graph obtained from GG^{*} by adding all black edges in HH^{\prime} which belong to the boundary of cellular 3-faces of HH^{\prime} and then subdividing each of these added edges. Let mm be the number of edges in HH^{\prime} that belong to the boundaries of cellular 3-faces of HH^{\prime}. Then

|V(G)|=|V(G)|+mand|E(G)|=|E(G)|+2m.|V(G^{**})|=|V(G^{*})|+m\quad\mbox{and}\quad|E(G^{**})|=|E(G^{*})|+2m. (11)

Because the edges of HH (and thus HH^{\prime}) are not crossed with the edges of GG (and thus GG^{*}), we observe that GG^{**} is also a simple and bipartite plane graph and has at least tt cellular 6-faces. Applying Lemma 3 to GG^{**} yields that

|E(G)|2|V(G)|4t.|E(G^{**})|\leq 2|V(G^{**})|-4-t.

Then, (11) implies that |E(G)|2|V(G)|4t|E(G^{*})|\leq 2|V(G^{*})|-4-t. This proves the claim. \Box

Claim 7: |E(G)|2|V(G)|+4|X|12t0/2|E(G)|\leq 2|V(G)|+4|X|-12-t_{0}/2.

Proof.  By (9), we have

|E(G)|=|E(G)|+|E(H)|+j{0,1,3}i=1tjei(j)(t1+3t3).|E(G)|=|E(G^{*})|+|E(H)|+\sum\limits_{j\in\{0,1,3\}}\sum\limits_{i=1}^{t_{j}}e^{(j)}_{i}-(t_{1}+3t_{3}).

Then, by Claims 4, 5 and 6,

|E(G)|\displaystyle|E(G)| \displaystyle\leq (2|V(G)|4t)+(4|X|8+t(3t0+2t1)/2)\displaystyle\Big{(}2|V(G^{*})|-4-t\Big{)}+\Big{(}4|X|-8+t-(3t_{0}+2t_{1})/2\Big{)}
+(2j{0,1,3}i=1tjyi(j)+(t0+2t1+3t3))(t1+3t3)\displaystyle+\Big{(}2\sum_{j\in\{0,1,3\}}\sum_{i=1}^{t_{j}}y^{(j)}_{i}+(t_{0}+2t_{1}+3t_{3})\Big{)}-(t_{1}+3t_{3})
=\displaystyle= 2|V(G)|+4|X|12t0/2by (10)\displaystyle 2|V(G)|+4|X|-12-t_{0}/2\hskip 142.26378pt\mbox{by }(\ref{G* vertex})
\displaystyle\leq 2|V(G)|+4|X|12.\displaystyle 2|V(G)|+4|X|-12.

\Box

Clearly, Claim 7 contradicts the assumption in (1). Hence Theorem 3 holds. \Box

6 Remarks

For any x3x\geq 3 and y6x12y\geq 6x-12, Czap, Przybylo and S̆krabuláková [6, Lemma 4] constructed a bipartite 11-planar graph GG with partite sets XX and YY such that |E(G)|=2|V(G)|+4|X|12|E(G)|=2|V(G)|+4|X|-12. Notice that the 11-planar drawing DD of this graph GG given in [6] has the following property:

(*) each vertex in XX is incident with crossed edges in DD.

The proof of Theorem 3 also yields that, if |E(G)|=2|V(G)|+4|X|12|E(G)|=2|V(G)|+4|X|-12 holds for a bipartite 11-planar graph GG with partite sets XX and YY, where 4|X||Y|4\leq|X|\leq|Y|, and DD is a 11-planar drawing of GG with the minimum number of crossings, then the graph HH^{\prime} introduced in Section 3 does not have isolated vertices, i.e., property (*) above holds.

Based on the above observations, we propose the following problem.

Problem 1

For any bipartite 11-planar graph GG with partite sets XX and YY, where 4|X||Y|4\leq|X|\leq|Y|, if |E(G)|=2|V(G)|+4|X|12|E(G)|=2|V(G)|+4|X|-12, does property (*) hold for every 11-planar drawing DD of GG with the minimum number of crossings?

From Claims 3 and 5 in the proof of Theorem 3, we can see that if HH^{\prime} does not have separating 22-cycles and |X>0|3|X_{>0}|\geq 3, where X>0X_{>0} is the set of non-isolated vertices in HH^{\prime} (i.e., the set of vertices in XX which are incident with crossed edges of DD), then |E(G)|2|V(G)|+4|X>0|12|E(G)|\leq 2|V(G)|+4|X_{>0}|-12 holds.

Problem 2

Let GG be a bipartite 11-planar graph with partite sets XX and YY, where 4|X||Y|4\leq|X|\leq|Y|. If DD is a 11-planar drawing of GG with the minimum number of crossings and |X>0|3|X_{>0}|\geq 3, where X>0X_{>0} is the set of vertices in XX which are incident with crossed edges of DD, does |E(G)|2|V(G)|+4|X>0|12|E(G)|\leq 2|V(G)|+4|X_{>0}|-12 hold?

Theorem 3 shows that |E(G)|2|V(G)|+4x12|E(G)|\leq 2|V(G)|+4x-12 holds for any bipartite 11-planar graph GG with bipartite sets of sizes xx and yy, where 2xy2\leq x\leq y. For any x3x\geq 3 and y6x12y\geq 6x-12, Czap, Przybylo and S̆krabuláková [6] constructed a bipartite 11-planar graph GG with bipartite sets of sizes xx and yy and |E(G)|=2|V(G)|+4x12|E(G)|=2|V(G)|+4x-12. Notice that these graphs constructed in [6] have minimum degree 33. By Theorem 1, any bipartite 11-planar graph of nn vertices has at most 3n83n-8 edges, implying that its minimum degree is at most 55. We wonder if the result in Theorem 3 can be improved for bipartite 11-planar graphs with higher minimum degrees or connectivity.

Problem 3

Let 4t54\leq t\leq 5 and GG be any bipartite 11-planar graph with partite sets XX and YY, where t|X||Y|t\leq|X|\leq|Y|. If GG is tt-connected (or δ(G)=t\delta(G)=t), does |E(G)|2|V(G)|+f(t)|X|+c|E(G)|\leq 2|V(G)|+f(t)|X|+c holds for some f(t)<4f(t)<4?

Let t2t\geq 2. A drawing of a graph is tt-planar if each of its edges is crossed at most tt times. If a graph has a tt-planar drawing, then it is tt-planar. Does Theorem 3 have an analogous result for bipartite 22-planar graphs?

Problem 4

Let GG be a bipartite 22-planar graph with partite sets XX and YY, where 2|X||Y|2\leq|X|\leq|Y|. Determine constants a,ba,b and cc such that |E(G)|a|V(G)|+b|X|+c|E(G)|\leq a|V(G)|+b|X|+c.

Lemma 1 gives an upper bound for the size of a bipartite graph GG with partite sets XX and YY, where |X|=3|X|=3, which has a 1-disc 𝒪X{\cal O}_{X} drawing. Can this result be generalized for such a bipartite graph without the condition that |X|=3|X|=3?

Problem 5

Let GG be a bipartite graph with partite sets XX and YY which has a 1-disc 𝒪X{\cal O}_{X} drawing. Is it true that |E(G)|2|Y|+5|X|/32|E(G)|\leq 2|Y|+5|X|/3-2?

References

  • [1]
  • [2] J. A. Bondy and U. S. R. Murty, Graph Theory, Grad. Texts in Math., vol.244, Springer, New York, 2008.
  • [3] H. Bodendiek, R. Schumacher and K. Wagner, Űber 1-optimale Graphen, Math. Nachr. 117 (1984), 323-339.
  • [4] J. Czap and D. Huda´\acute{a}k, 1-planarity of complete multipartite graphs, Discrete Appl. Math. 160 (2012), 505-512.
  • [5] J. Czap and D. Huda´\acute{a}k and T. Madras, Joins of 1-planar graphs, Acta Mathematica Sinica, English Series 30(11) (2014), 1867-1876.
  • [6] J. Czap, J. Przybylo, and E. S̆krabuláková, On an extremal problem in the class of bipartite 1-planar graphs, Discuss. Math., Graph Theory 36 (2016), 141-151.
  • [7] I. Fabrici and T. Madars, The structure of 1-planar graphs, Discrete Math. 307 (2007), 854-865.
  • [8] D.V. Karpov, Upper bound on the number of edges of an almost planar bipartite graph, J. Math. Sci. 196 (2014), 737-706.
  • [9] S. G. Kobourov, G. Liotta and F. Montecchiani, An annotated bibliography on 1-planarity, Comput. Sci. Rev. 25 (2017), 49-67.
  • [10] D.J. Kleitman, The crossing number of K5,nK_{5,n}, J. Combin. Theory 9 (1970), 315-323.
  • [11] J. Pach and G. Tóth, Graphs drawn with few crossings per edge, Combinatorica 17(3) (1997), 427-439.
  • [12] G. Ringel, Ein Sechsfarbenproblem auf der Kugel, Abh. Math. Sem. Univ. Hamburg 29 (1965) 107-117.
  • [13] E. Sopena, personal communication, Seventh Cracow Conference in Graph Theory, Rytro (2014).
  • [14] X. Zhang, Drawing complete multipartite graphs on the plane with restrictions on crossings, Acta Mathematica Sinica, English Series 30 (12) (2014), 2045-2053.