Skip to main content
Log in

2-Distance coloring of planar graphs with girth 5

  • Published:
Journal of Combinatorial Optimization Aims and scope Submit manuscript

Abstract

A vertex coloring is said to be 2-distance if any two distinct vertices of distance at most 2 receive different colors. Let G be a planar graph with girth at least 5. In this paper, we prove that G admits a 2-distance coloring with at most \(\Delta (G)+3\) colors if \(\Delta (G)\ge 339\).

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9

Similar content being viewed by others

References

  • Bondy JA, Murty USR (1976) Graph theory with applications. Macmillan Ltd. Press, New York

    Book  MATH  Google Scholar 

  • Borodin OV (2013) Colorings of plane graphs: a survey. Discrete Math. 313:517–539

    Article  MathSciNet  MATH  Google Scholar 

  • Borodin OV, Ivanova AO (2009) 2-Distance \(\Delta +2\)-coloring of planar graphs with girth six and \(\Delta \ge 18\). Discrete Math. 309:6496–6502

    Article  MathSciNet  MATH  Google Scholar 

  • Borodin OV, Ivanova AO (2011a) Injective (\(\Delta +1\))-coloring of planar graphs with girth six. Sibirsk. Mat. Zh. 52:30–38 (in Russian)

    Article  MathSciNet  MATH  Google Scholar 

  • Borodin OV, Ivanova AO (2011b) List injective colorings of planar graphs. Discrte Math. 311:154–165

    Article  MathSciNet  MATH  Google Scholar 

  • Borodin OV, Broersma H, Glebov H, van den Heuvel J (2002) Stars and bunches in planar graphs. General planar graphs and colourings. Part II. CDAM Researches Report

  • Borodin OV, Ivanova AO, Neustroeva TK (2004a) 2-Distance coloring of sparse plane graphs. Sib. Èlektron. Mat. Izv. 1:76–90 (in Russian)

    MathSciNet  MATH  Google Scholar 

  • Borodin OV, Glebov AN, Ivanova AO, Neustroeva TK, Tashkinov VA (2004b) Sufficient conditions for the 2-distance \(\Delta +1\)-colorability of plane graphs. Sib. Èlektron. Mat. Izv. 1:129–141 (in Russian)

    MathSciNet  MATH  Google Scholar 

  • Bu Y, Zhu X (2012) An optimal square coloring of planar graphs. J. Comb. Optim. 24:580–592

    Article  MathSciNet  MATH  Google Scholar 

  • Cranston D, Kim S, Yu G (2010) Injective coloring of sparse graphs. Discrete Math. 310:2965–2973

    Article  MathSciNet  MATH  Google Scholar 

  • Cranston D, Kim S, Yu G (2011) Injective coloring of graphs with low average degree. Algorithmica 60:553–568

    Article  MathSciNet  MATH  Google Scholar 

  • Dong W, Lin W (2013) Injective coloring of planar graphs with girth 6. Discrete Math. 313:1302–1311

    Article  MathSciNet  MATH  Google Scholar 

  • Dong W, Lin W (2014) Injective coloring of plane graphs with girth 5. Discrete Math. 315–316:120–127

    Article  MathSciNet  MATH  Google Scholar 

  • Dong W, Lin W (2016) An improved bound on 2-distance coloring plane graph with girth 5. J. Comb. Optim. 32:645–655

    Article  MathSciNet  MATH  Google Scholar 

  • Dong W, Lin W (2017) On 2-distance coloring of plane graphs with girth 5. Discrete Appl. Math. 217:495–505

    Article  MathSciNet  MATH  Google Scholar 

  • Doyon A, Hahn G, Raspaud A (2010) Some bounds on the injective chromatic number of graphs. Discrete Math. 310:585–590

    Article  MathSciNet  MATH  Google Scholar 

  • Dvořàk Z, Kràl D, Nejedlỳ P, Škrekovski R (2008) Coloring squares of planar graphs with girth six. Eur J Comb 29:838–849

    Article  MathSciNet  MATH  Google Scholar 

  • Hahn G, Kratochvíl J, Širáň J, Sotteau D (2012) On the injective chromatic number of graphs. Discrete Math 256:179–192

    Article  MathSciNet  MATH  Google Scholar 

  • Lužar B, S̆krekovski R, Tancer M (2009) Injective colorings of planar graphs with few colors. Discrete Math 309:5636–5649

    Article  MathSciNet  MATH  Google Scholar 

  • Molloy M, Salavatipour MR (2005) A bound on the chromatic number of the square of a planar graph. J Comb Theory Ser B 94:189–213

    Article  MathSciNet  MATH  Google Scholar 

  • van den Heuvel J, McGuinness S (2003) Coloring of the square of planar graph. J Graph Theory 42:110–124

    Article  MathSciNet  MATH  Google Scholar 

  • Wang W, Lih K (2003) Labeling planar graphs with conditions on girth and distance two. SIAM J Discrete Math 17:264–275

    Article  MathSciNet  MATH  Google Scholar 

  • Wegner G (1977) Graphs with given diameter and a coloring problem. Technical Report, University of Dortmund, Germany

Download references

Acknowledgements

The authors thank the referees for their valuable suggestions which helped to improve the presentation.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Wei Dong.

Additional information

Supported by Natural Science Foundation of Jiangsu Province of China (No. BK20140089) and NSFC Nos. 11331003 and 11571180.

Appendix: The proofs of the case that v is of type P7 or P8

Appendix: The proofs of the case that v is of type P7 or P8

Suppose that v is of type P7 (i.e., \(d(v)=4\), \(n_2^l(v)=2\) and \(n_2^h(v)=1\)). Let \(N(v)=\{x,y,z,u\}\) in clockwise order with \(d(x)=d(y)=d(z)=2\). By Lemma 2.3(8), \(d(u)\ge \Delta \). Let \(x_1,y_1\) and \(z_1\) be the other neighbor of xy and z, and let \(f_1,f_2,f_3\) and \(f_4\) be the faces with yvxzvyuvz and xvu in their boundaries, respectively. If \(d(f_i)\ge 6\) for some \(1\le i\le 4\), then \(ch(f_i\rightarrow v)\ge \frac{1}{2}\) since \(n_b(f_i)\le d(f_i)-2\). Hence, \(w^*(v)\ge w^\prime (v)+\frac{1}{2}\ge 0\). Otherwise, we suppose that \(d(f_i)=5\) for all \(1\le i\le 4\). Let \(f_1=[yvxx_1y_1]\), \(f_2=[zvyy_1z_1]\), \(f_3=[uvzz_1u_1]\) and \(f_4=[xvuu_2x_1]\).

We have two possibilities.

Case P71 y is the heavy 2-neighbor of v. It follows that \(d(y_1)\ge \Delta -1\). If \(d(x_1)\ge 10\), then \(x_1y_1\) is a heavy edge, and \(ch(x_1\rightarrow f_1)\ge 1\) and \(ch(y_1\rightarrow f_1)\ge \frac{3-\varepsilon }{2}\) by \(R_{21}\), which imply \(ch(f_1\rightarrow v)\ge 1+\frac{3-\varepsilon }{2}\) (as \(n_b(f_1)=1\)), and so \(w^*(v)\ge w'(v)+1+\frac{3-\varepsilon }{2}\ge 0\). Otherwise, we suppose that \(3\le d(x_1), d(z_1)\le 9\) by symmetry. By Lemma 2.2, both \(x_1\) and \(z_1\) have \(p^+\)-neighbors. With the similar reason, we also suppose that \(\max \{d(u_1), d(u_2)\}\le 9\).

Let us calculate \(ch(\{f_1, f_2, f_3, f_4\}\rightarrow v)\).

If \(d(x_1)\ge 5\), then \(w'(x_1)\ge 3d(x_1)-10+3-\varepsilon -(1+\varepsilon )-2(d(x_1)-2)\ge d(x_1)-4-2\varepsilon \) by \(R_{11}\) and \(R_{15}\) (noticing that \(u_2\) is heavy if \(d(u_2)=2\)), which implies \(ch(x_1\rightarrow f_1)\ge \frac{1-2\varepsilon }{5}\). So, \(ch(\{f_1, f_2, f_3, f_4\}\rightarrow v)\ge \frac{1-2\varepsilon }{5}\) (as \(n_b(f_4)=1\)), and \(w^*(v)\ge -2\varepsilon +\frac{1-2\varepsilon }{5}\ge 0\). Suppose so that \(d(x_1)\le 4\).

If \(d(x_1)=3\), then \(d(u_2)\ge 3\) by Lemma 2.3(3), and \(u_2\) is not a 3(1)-vertex with \(n_2^l(u_2)=1\) by Lemma 2.6(1). Now, after the first discharging procedure, \(w'(u_2)\ge -1+3-\varepsilon -(1+\varepsilon )-\varepsilon \ge 1-3\varepsilon \) if \(d(u_2)=3\) (by \(R_{11}\)-\(R_{13}\) and \(R_{15}\)), \(w'(u_2)\ge 3d(u_2)-10+3-\varepsilon -2(d(u_2)-2)-\varepsilon \ge d(u_2)-3-2\varepsilon \) (by \(R_{11}\), \(R_{14}\) and \(R_{15}\)) whenever \(d(u_2)\ge 4\). Hence \(ch(u_2\rightarrow f_4)\ge \frac{1-2\varepsilon }{4}\), which implies \(ch(\{f_1, f_2, f_3, f_4\}\rightarrow v)\ge \frac{1-2\varepsilon }{4}\) (since \(n_b(f_4)=1\)), and so \(w^*(v)\ge -2\varepsilon +\frac{1-2\varepsilon }{4}\ge 0\).

Finally, we suppose that \(d(x_1)=4\). By Lemma 2.3(8), \(u_2\) is not a light 2-vertex. Therefore, \(x_1\) is not a 4-vertex with \(n_2^l(x_1)=2\) and \(n_2^h(x_1)=1\) by Lemma 2.6(2), and \(w'(x_1)\ge 2+3-\varepsilon -4-\varepsilon \ge 1-2\varepsilon \) whenever \(n_2^l(x_1)\)=2 (noticing that \(x_1\) transfers at most \(\varepsilon \) to \(u_2\) in this situation) or \(w'(x_1)\ge 2+3-\varepsilon -2-2(1+\varepsilon )\ge 1-3\varepsilon \) whenever \(n_2^l(x_1)\le 1\). It follows that \(ch(x_1\rightarrow f_1)\ge \frac{1-3\varepsilon }{4}\), and \(ch(f_1\rightarrow v)\ge \frac{1-3\varepsilon }{4}\) (\(n_b(f_1)=1\)), and so \(w^*(v)\ge -2\varepsilon +\frac{1-3\varepsilon }{4}\ge 0\).

Case P72 x is the heavy 2-neighbor of v. It follows that \(d(x_1)\ge p\). If \(d(u_2)\ge 10\), then \(u_2x_1\) and \(uu_2\) are both heavy edges, and \(f_4\) receives at least \(2+3-\varepsilon \) by \(R_{21}\), and so \(w^*(v)\ge -2\varepsilon +5-\varepsilon \ge 0\). Otherwise, we have \(d(u_2)\le 9\). It is easy to check that \(w'(u_2)\ge 3d(u_2)-10+2(3-\varepsilon )-2(d(u_2)-2)\ge d(u_2)-2\varepsilon \) since \(u_2\) receives at least \(3-\varepsilon \) from u and \(x_1\), respectively. By \(R_{22}\), \(ch(u_2\rightarrow f_4)\ge \frac{2-2\varepsilon }{2}\). Hence by \(R_{23}\), \(w^*(v)\ge -2\varepsilon +ch( f_4\rightarrow v)\ge -2\varepsilon +\frac{2-2\varepsilon }{2}\ge 0\) since \(n_b(f_4)=1\).

Suppose that v is of type P8 (i.e., v is a 4-vertex with \(n_2^l(v)=3\)). Let \(N(v)=\{x,y,z,u\}\) with xyz being light 2-vertices. By Lemma 2.3(8), \(d(u)\ge \Delta \). For convenience, let \(x_1,y_1\) and \(z_1\) be the other neighbor of xy and z, respectively. By the definition of light 2-vertex, max \(\{d(x_1),d(y_1),d(z_1)\}\le \Delta -2\). For convenience, we assume that \(f_1,f_2,f_3\) and \(f_4\) are the four faces incident with v, bounded by yvxzvyuvz and xvu, respectively. If v is incident with at least two \(6^+\)-faces, then by Claim 3.4 and \(R_{23}\), \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+\frac{1}{2}\ge 0\) or \(w^*(v)\ge -1-\varepsilon +\frac{1}{2}\times 3\ge 0\). So we assume that v is incident with at most one \(6^+\)-face. We have three possibilities.

Case P81 \(d(f_2)\ge 6\). Let \(f_1=[yvxx_1y_1]\), \(f_3=[uvzz_1u_1]\) and \(f_4=[xvuu_2x_1]\) with \(u_1,u_2\in N(u)\). We will calculate \(ch(\{f_1, f_2, f_3, f_4\}\rightarrow v)\). Notice that \(ch(f_2\rightarrow v)\ge \frac{2}{3}\) by Claim 3.4.

If \(d(x_1)\ge 10\), then \(w^\prime (u_2)\ge 3d(u_2)-10+2+3-\varepsilon -2(d(u_2)-2)\ge d(u_2)-1-\varepsilon \) since \(u_2\) receives at least 2 from \(x_1\) and receives \(3-\varepsilon \) from u. By \(R_{22}\), \(ch(u_2\rightarrow f_4)\ge \frac{1-\varepsilon }{2}\) and \(ch(f_4\rightarrow v)\ge \frac{1-\varepsilon }{2}\) since \(n_b(f_4)=1\). Hence, we have \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+\frac{1-\varepsilon }{2}\ge 0\).

Otherwise, we may assume that \(3\le d(x_1)\le 9\). By Lemma 2.2, \(x_1\) is adjacent to a \(p^+\)-vertex. We claim that \(d(y_1)\le p-1\). ( Otherwise, \(w'(y)\ge -4+2+3-\varepsilon \ge 1-\varepsilon \) by \(R_{11}\) and \(R_{15}\) in the first step. By \(R_{22}\), \(ch(y\rightarrow v)\ge 1-\varepsilon \). Hence, \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+1-\varepsilon \ge 0\).) If \(d(u_2)\ge 10\), then \(uu_2\) is a heavy edge. By \(R_{21}\), \(ch(u\rightarrow f_4)\ge \frac{3-\varepsilon }{2}\) and \(ch(u_2\rightarrow f_4)\ge 1\). We have \(ch(f_4\rightarrow v)\ge \frac{3-\varepsilon }{4}+\frac{1}{2}\) since \(n_b(f_4)\le 2\). It follows that \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+\frac{3-\varepsilon }{4}+\frac{1}{2}\ge 0\). Otherwise, we assume that \(2\le d(u_2)\le 9\).

If \(d(u_2)=2\), then by Lemma 2.3(4), \(d(x_1)\ge 4\). If \(d(x_1)=4\), \(w^\prime (x_1)\ge 2+3-\varepsilon -2-2(1+\varepsilon )\ge 1-3\varepsilon \) by Lemma 2.3(8). If \(d(x_1)\ge 5\), then \(w^\prime (x_1)\ge 3d(x_1)-10+3-\varepsilon -(1+\varepsilon )-2(d(x_1)-2)\ge d(x_1)-4-2\varepsilon \) since \(x_1\) receives at least \(3-\varepsilon \) from its \(p^+\)-neighbors and transfers at most \(1+\varepsilon \) to \(u_2\). We have \(ch(x_1\rightarrow \{f_1, f_4\})\ge \frac{1-2\varepsilon }{5}\times 2\). Hence, \(ch(\{f_1, f_4\}\rightarrow v)\ge \frac{1-2\varepsilon }{5}\times 2\) since \(n_b(f_1)=1\) and \(n_b(f_4)=1\). By \(R_{23}\), \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+\frac{1-2\varepsilon }{5}\times 2\ge 0\).

Finally, we assume that \(d(u_2)\ge 3\). If \(d(x_1)\ge 4\), then \(w^\prime (x_1)\ge 3d(x_1)-10+3-\varepsilon -2(d(x_1)-2)-\varepsilon \ge d(x_1)-3-2\varepsilon \) since \(x_1\) receives at least \(3-\varepsilon \) from its \(p^+\)-neighbors and transfers at most \( \varepsilon \) to \(u_2\). We have \(ch(x_1\rightarrow \{f_1, f_4\})\ge \frac{1-2\varepsilon }{2}\). Hence \(ch(\{f_1, f_4\}\rightarrow v)\ge \frac{1-2\varepsilon }{2}\) since \(n_b(f_1)=1\) and \(n_b(f_4)=1\). \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+\frac{1-2\varepsilon }{2}\ge 0\). Otherwise, if \(d(x_1)=3\), then \(d(u_2)\ge 5\) by Lemma 2.7. We have \(w^\prime (u_2)\ge 3d(u_2)-10+3-\varepsilon -\varepsilon -2(d(u_2)-2)\ge d(u_2)-3-2\varepsilon \) since \(u_2\) receives at least \(3-\varepsilon \) from u and transfers at most \(\varepsilon \) to \(x_1\). Hence, \(ch(u_2\rightarrow f_4)\ge \frac{2-2\varepsilon }{5}\). It follows that \(w^*(v)\ge -1-\varepsilon +\frac{2}{3}+\frac{2-2\varepsilon }{5}\ge 0\).

Case P82 \(d(f_3)\ge 6\). For convenience, let \(f_1=[yvxx_1y_1]\), \(f_2=[zvyy_1z_1]\) and \(f_4=[x_1xvuu_1]\) be the three 5-faces incident with v. By Claim 3.4 and \(R_{23}\), \(ch(f_3\rightarrow v)\ge \frac{2}{3}\). We consider \(ch(u_2\rightarrow f_4)\) and \(ch(x_1\rightarrow \{f_1, f_4\})\). This case is quite similar to that of Case P81, so we omit it.

Case P83 \(d(f_i)=5\) for all \(1\le i\le 4\). Without loss of generality, let \(f_1=[yvxx_1y_1]\), \(f_2=[zvyy_1z_1]\), \(f_3=[uvzz_1u_1]\) and \(f_4=[xvuu_2x_1]\) with \(u_1,u_2\in N(u)\). If \(d(u_2)\ge 10\), then \(uu_2\) is a heavy edge, \(ch(\{u,u_2\}\rightarrow f_4)\ge 1+\frac{3-\varepsilon }{2}\) and \(ch(f_4\rightarrow v)\ge \frac{1}{2}+\frac{3-\varepsilon }{4}\) since \(n_b(f_4)\le 2\). It follows that \(w^*(v)\ge -1-\varepsilon +\frac{1}{2}+\frac{3-\varepsilon }{4}\ge 0\).

So we may assume that \(2\le d(u_1)\le 9\) and \(2\le d(u_2)\le 9\) by symmetry. In the following, we will calculate the charge transferred from \(x_1,z_1,y_1,u_1,u_2\) to \(f_1,f_2,f_3\) and \(f_4\).

Subcase P831 \(d(x_1)\ge 10\) and \(d(z_1)\ge 10\). If \(d(y_1)\ge 10\), then both \(x_1y_1\) and \(y_1z_1\) are heavy edges. Hence, \(ch(\{x_1,y_1,z_1\}\rightarrow \{f_1, f_2\})\ge 4\). It follows that \(w^*(v)\ge -1-\varepsilon +4\ge 0\) since \(n_b(f_1)=n_b(f_2)=1\). Otherwise, we may assume that \(3\le d(y_1)\le 9\). By Lemma 2.2, \(y_1\) is adjacent to at least one \(p^+\)-vertex. It is easy to calculate that \(w^\prime (u_2)\ge 3d(u_2)-10+2+3-\varepsilon -2(d(u_2)-2)\ge d(u_2)-1-\varepsilon \) by \(R_{11}\) and \(R_{15}\). By \(R_{22}\), \(ch(u_2\rightarrow f_4)\ge \frac{1-\varepsilon }{2}\) and \(ch(f_4\rightarrow v)\ge \frac{1-\varepsilon }{2}\) since \(n_b(f_4)=1\). By symmetry, \(ch(f_3\rightarrow v)\ge \frac{1-\varepsilon }{2}\).

Moreover, by \(R_{11}\) and \(R_{15}\), \(w^\prime (y_1)\ge 3d(y_1)-10+4-2(d(y_1)-2))\ge d(y_1)-2\), which shows that \(ch(y_1\rightarrow \{f_1, f_2\})\ge \frac{2}{3}\). Again, since \(n_b(f_1)=n_b(f_2)=1\), \(ch(\{f_1,f_2\}\rightarrow v)\ge \frac{2}{3}\). Totally, \(ch(\{f_1, f_2, f_3, f_4\}\rightarrow v)\ge \frac{1-\varepsilon }{2}+\frac{1-\varepsilon }{2}+\frac{2}{3}\). It follows that \(w^*(v)\ge -1-\varepsilon +(1-\varepsilon )+\frac{2}{3}\ge 0\).

Subcase P832 \(d(x_1)\ge 10\) and \(d(z_1)\le 9\). Similar to that of above, if \(d(y_1)\ge 10\), then \(x_1y_1\) is a heavy edge. Hence, \(ch(\{x_1,y_1\}\rightarrow f_1)\ge 2\). It follows that \(w^*(v)\ge -1-\varepsilon +2\ge 0\) since \(n_b(f_1)=1\). Otherwise, we may assume that \(3\le d(y_1)\le 9\).

If \(10\le d(x_1)\le p-1\), then \(d(y_1)\ge 4\) and \(y_1\) is adjacent to a \(p^+\)-vertex by Lemma 2.2. It is easy to check that \(w^\prime (y_1)\ge 3d(y_1)-10+2+3-\varepsilon -\varepsilon -2(d(y_1)-3))\ge d(y_1)+1-2\varepsilon \) since \(y_1\) receives 2 from \(x_1\), receives \(3-\varepsilon \) from it \(p^+\)-neighbors and transfers at most \(\varepsilon \) to \(z_1\). By \(R_{22}\), \(ch(y_1\rightarrow f_1)\ge \frac{5-2\varepsilon }{4}\) and \(ch(y_1\rightarrow f_2)\ge \frac{5-2\varepsilon }{4}\). It follows that \(ch(f_1\cup f_2\rightarrow v)\ge \frac{5-2\varepsilon }{4}\times 2\) as \(n_b(f_1)=n_b(f_2)=1\). By \(R_{23}\), \(w^*(v)\ge -1-\varepsilon +\frac{5-2\varepsilon }{2}\ge 0\).

Otherwise, we assume that \(d(x_1)\ge p\). It is easy to calculate that \(w^\prime (u_2)\ge 3d(u_2)-10+3-\varepsilon +3-\varepsilon -2(d(u_2)-2)\ge d(u_2)-2\varepsilon \) since \(u_2\) receives \(3-\varepsilon \) from \(x_1\) and u, respectively. By \(R_{22}\), \(ch(u_2\rightarrow f_4)\ge 1-\varepsilon \) and \(ch(f_4\rightarrow v)\ge {1-\varepsilon }\) since \(n_b(f_4)=1\). On the other hand, \(w^\prime (x)\ge -4+2+3-\varepsilon \ge 1-\varepsilon \). By \(R_{22}\), \(ch(x\rightarrow v)\ge 1-\varepsilon \). Totally, \(w^*(v)\ge -1-\varepsilon +(1-\varepsilon )+(1-\varepsilon )\ge 0\).

Subcase P833 \(d(x_1)\le 9\) and \(d(z_1)\le 9\).

Notice that min\(\{d(x_1),d(y_1),d(z_1)\}\ge 4\) in this situation by Lemma 2.2. First we assume that \(d(y_1)\le p-1\). It is not difficult to check that \(w^\prime (z_1)\ge 3d(z_1)-10+3-\varepsilon -(1+\varepsilon )-\varepsilon -2(d(z_1)-3)\ge d(z_1)-2-3\varepsilon \) (noticing that \(z_1\) receives \(3-\varepsilon \) from its \(p^+\)-neighbors and transfers at most \(1+\varepsilon \), \(\varepsilon \) to \(u_1\) and \(y_1\), respectively). It follows that \(ch(z_1\rightarrow \{f_2, f_3\})\ge \frac{2-3\varepsilon }{4}\times 2\) and \(ch(\{f_2, f_3\}\rightarrow v)\ge \frac{2-3\varepsilon }{4}\times 2\) since \(n_b(f_2)=n_b(f_3)=1\). By symmetry, \(ch(x_1\rightarrow \{f_1, f_4\})\ge \frac{2-3\varepsilon }{4}\times 2\). We have \(w^*(v)\ge -1-\varepsilon +\frac{2-3\varepsilon }{4}\times 4\ge 0\).

Otherwise, we assume that \(d(y_1)\ge p\). Notice that \(w^\prime (y)\ge -4+2+3-\varepsilon \ge 1-\varepsilon \) by \(R_{11}\) and \(R_{15}\). Hence, \(ch(y\rightarrow v)\ge 1-\varepsilon \) by \(R_{22}\).

Besides, we also calculate the charge transferred from \(x_1\) to it incident faces.

If \(d(x_1)\ge 5\), then \(w^\prime (x_1)\ge 3d(x_1)-10+3-\varepsilon -(1+\varepsilon )-2(d(x_1)-2)\ge d(x_1)-4 -\varepsilon \) since \(x_1\) receives \(3-\varepsilon \) from \(y_1\) and transfers at most \(1+\varepsilon \) to \(u_2\). It follows that \(ch(x_1\rightarrow \{f_1, f_4\})\ge \frac{1-\varepsilon }{5}\times 2\), which induces that \(w^*(v)\ge -1-\varepsilon +1-\varepsilon +\frac{1-\varepsilon }{5}\times 2\ge 0\) since \(n_b(f_1)=n_b(f_4)=1\).

If \(d(x_1)=4\), then by Lemma 2.7(2), \(n_2(x_1)\le 2\). Hence, \(w^\prime (x_1)\ge 2+3-\varepsilon -4-\frac{1}{3}\ge \frac{2}{3}-\varepsilon \). It follows that \(ch(x_1\rightarrow \{f_1, f_4\})\ge \frac{2-3\varepsilon }{12}\times 2\), which induces that \(w^*(v)\ge -1-\varepsilon +1-\varepsilon +\frac{2-3\varepsilon }{12}\times 2\ge 0\).

Finally, if \(d(x_1)=3\), then by Lemma 2.7(1), \(d(u_2)\ge 5\). We consider \(u_2\). It is easy to check that \(w^\prime (u_2)\ge 3d(u_2)-10+3-\varepsilon -\varepsilon -2(d(u_2)-2)\ge d(u_2)-3-2\varepsilon \) since \(u_2\) receives \(3-\varepsilon \) from u and transfers at most \(\varepsilon \) to \(x_1\). Hence \(ch(u_2\rightarrow f_4)\ge \frac{2-2\varepsilon }{5}\) and \(ch(f_4\rightarrow v)\ge \frac{2-2\varepsilon }{5} \), which induces that \(w^*(v)\ge -1-\varepsilon +1-\varepsilon +\frac{2-2\varepsilon }{5}\ge 0\).

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Dong, W., Xu, B. 2-Distance coloring of planar graphs with girth 5. J Comb Optim 34, 1302–1322 (2017). https://doi.org/10.1007/s10878-017-0148-7

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10878-017-0148-7

Keywords

Mathematics Subject Classification

Navigation