Radio--Labeling of Cycles for Large ††thanks: Research is partially supported by the National Science Foundation grant DMS-1600778 and NASA NASA MIRO grant NX15AQ06A.
Abstract
Let be a simple connected graph. For any two vertices and , let denote the distance between and in . A radio--labeling of for a fixed positive integer is a function which assigns to each vertex a non-negative integer label such that for every two vertices and in , . The span of is the difference between the largest and smallest labels of . The radio--number of a graph , denoted by , is the smallest span among all radio--labelings admitted by . A cycle has diameter . In this paper, we combine a lower bound approach with cyclic group structure to determine the value of for . For , we obtain the values of when and have the same parity, and prove partial results when and have different parities. Our results extend the known values of and shown by Liu and Zhu [15], and by Karst, Langowitz, Oehrlein and Troxell [10], respectively.
Keywords: Frequency assignment problem; radio labelling; radio--labeling, radio--number; cyclic groups.
1 Introduction
Radio--labeling is motivated by the channel assignment problem which was first introduced by Hale [8]. In the channel assignment problem, we must both avoid signal interference between radio stations and minimize the range of frequencies used. Radio--labeling models this problem with a graph where each station or transmitter is a vertex and neighboring stations are connected by an edge. The distance between two stations is then the usual distance between vertices on a graph, and a station’s assigned frequency is represented by a non-negative integer label of the corresponding vertex. If two stations are close enough for their signals to interfere, they must be assigned frequencies whose difference is determined by their proximity: The closer the stations are geographically, the larger their separation in assigned frequencies must be. The goal is to assign frequencies to all stations in a way that minimizes the range of the frequencies used while still avoiding interference.
Let be a graph. The distance between two vertices and , denoted by , is the length of a shortest path between and . The diameter of is the largest distance between any two vertices in , which we denote by or sometimes when the graph is clear in the context.
For a positive integer , a function is a radio--labeling of a graph , if for all , ,
Sometimes we say is a labeling instead of radio--labeling or omit mentioning the labeled graph if it is clear from the context. The span of a radio--labeling of is defined as We follow the convention of labeling at least one vertex , so the span of any radio--labeling is the largest label assigned to a vertex .
The radio--number of is the minimum span of a radio--labeling admitted by . We call a radio--labeling of a graph optimal if the span is equal to the radio--number. A special case for the radio--labeling of a graph is when , in which a radio--labeling is also known as radio labeling and the radio number of is denoted by , where .
The notion of radio labeling was introduced by Chartrand, Erwin and Zhang [4]. The radio number for special families of graphs has been studied widely in the literature. Liu and Zhu [15] determined the radio number of cycles and paths. A general lower bound of the radio number of trees was given by Liu [13], and then applied by several authors to special families of trees (cf. [12, 7, 2, 14]). Khennoufa and Togni [11] studied the radio number for hypercubes by using generalized binary Gray codes. Zhou [20, 21, 22] and Martinez et al. [16] investigated the radio number for Cayley graphs and generalized prism graphs, respectively. Niedzialomski [18] studied radio graceful graphs (where admits a surjective radio labeling from to ) and showed that the Cartesian product of copies of a complete graph is radio graceful for certain , providing infinitely many examples of radio graceful graphs of arbitrary diameter. For two positive integers , the toroidal grid is the Cartesian product of cycles and , . Morris et al. [17] determined , and Saha and Panigrahi [19] determined when . Bantva and Liu [1] considered the radio number for block graphs and line graphs of trees.
When , radio--labeling is equivalent to vertex coloring. Radio--labelings of graphs for other values of have have also been studied extensively. There are many known results when (cf. [6, 3]), this case is called -labeling, which can be generalized to -labeling. Recently, Chavez et al. studied optimal radio--labeling for trees [5].
For cycles with , the value of was completely determined in [15]. Also, when has been determined in [9], where bounds are given for the case when . Karst, Langowitz, Oehrlein, and Troxell in [10] studied radio--labeling when and determined the value of when .
In this paper we investigate the values of for all . In particular, we determine the values of for , and for with (mod 2). For and (mod 2), we establish lower and upper bounds and show these bounds are sharp for some values of and . A major tool we use is a combination of a lower bound approach (see 3) with some basic properties of finite cyclic groups.
2 Preliminary and Notations
In what follows, we let denote the set of integers, the set of natural numbers and . For , such that , we define . For , we let denote the cyclic group of order . For , let be the subgroup of generated by . That is, (mod ). Let denote the cyclic quotient group of modulo , which is defined on the cosets for .
We regard as a Cayley graph with vertex set and edges generated by . That is, where and are adjacent if (mod ). In the proofs we will often make use of an additive structure on which is addition modular . The modest amount of group theory mentioned in this paper can be found in any introductory textbook on abstract algebra.
By definition, when , every radio--labeling of is injective. So we can order the vertices of such that the labeling is strictly increasing. Thus, given a radio--labeling of with , we denote the vertices of , for such that if , then for For , we define the label gap and the distance gap respectively by
These notations are used throughout the paper. By definition, .
Definition 1.
When defining a labeling, we first specify an ordering of the vertices of by defining a jumping (or jump) sequence with , . This jumping sequence gives the “clockwise distance" of from , namely, if , then (mod ).
Definition 2.
Let , , . For , we define the jump sequence generated by , , with and , where the indices of the jumps are taken modulo and the sums modulo . We call a periodic jump sequence with jumps. We call the range of the set generated by , and denote it as .
Here is an example of Definition 2.
Example 1.
Let and , so . Then , And
Notice that in 1, the jump sequence does not determine a permutation and so does not define a labeling for . In practice, we will define a jump sequence and then prove that it determines a permutation of the vertices of .
Lemma 1.
Fix . Let , , and . Then
Proof.
Note that in , . Hence . If , then . Thus, , so .
If , we look at the first vertices obtained from the jump sequence. The set of even and odd indexed terms, respectively, are
Since and , we have that and belong to different cosets of . Hence, and . Thus, . ∎
We will frequently use the following lemma in our proofs.
Lemma 2.
Let be even and . Let be the cyclic subgroup of generated by . Then
Proof.
Let , , where , , and both and are odd. If , then . It follows that and so .
If , then and so . It follows that . ∎
For any and , the following definition and notation will be used throughout the paper:
(1) |
The next result of Karst, Langowitz, Oehrlein, and Troxell [10] plays a critical role in our proofs. For completeness we include a proof.
Proposition 3.
[10] For and where and , we have the following lower bounds:
Proof.
Assume is a radio--labeling of . For the following hold:
-
1.
,
-
2.
,
-
3.
Adding these three inequalities we get
Since , we have
Hence, we obtain the lower bounds for the radio--number of :
The result follows. ∎
Denote the lower bounds of from 3 by :
(2) |
The next result from [10] greatly reduces the number of inequalities to verify when determining if a map is a radio--labeling.
Proposition 4.
[10] If , then is a radio--labeling of if and only if the following hold:
-
1.
-
2.
.
We do not include a proof of this result, however, the proof of 5 in the next section follows similarly.
3 Exact Values for
In this section, we completely determine the radio--number for when (Theorem 6) and when wile and have the same parity (Theorem 8).
For , we make use of the following lemma, which is a stronger version of Proposition 4 for the special case when .
Lemma 5.
For , a function is a radio--labeling of if and only if for all .
Proof.
The forward direction follows from the definition of radio--labeling. Let and satisfy , for all . By Proposition 4 we only need to show that is satisfied. Let . Then and Combining these two inequalities,
(3) |
Since ( and ) and , we have
Hence the result follows. ∎
Theorem 6.
Let . Then
Proof.
Suppose is a radio--labeling for . From Equation (3):
Let . If is odd, then
If is even, the last vertex cannot be paired with others and so,
Therefore, the lower bounds are established.
Next we find a labeling which meets the lower bounds. Suppose for . Let the distance gap be for all . Since and , if we let , then determines a permutation of . Define the label gap by . Then for all . Therefore, is a valid radio--labeling of and meets the desired lower bound.
Now consider the case when is even. That is, for , and . For we define the jumping and labeling sequences respectively by:
It is easy to see that the jump sequence gives a permutation of , and it holds that for all . By 5, is a radio--labeling of with the desired span. ∎
Lemma 7.
Let be a radio--labeling of , where and have the same parity. For any , if , then .
Proof.
Suppose . Because and have the same parity, by the definition of radio--labeling, we have
Hence, . Combining this with the fact that we get . Moreover, because , we obtain . Therefore, the result follows. ∎
7 shows that when and have the same parity, in order to keep the -function value for any three consecutive vertices, , the distance between and must be either or .
Theorem 8.
If and have the same parity and , then .
Proof.
Throughout the proof (in fact, throughout the paper), in every defined labeling, holds for all considered, which implies that Also, in every defined jump sequence, no jump will exceed the diameter of the graph, so that
Case 1. Both and are even. Denote and , where and . Let and . Define the jump sequence (or distance gap) by:
Claim: The jump sequence gives a permutation.
Proof of Claim) Recall from Section 1 that . If , then our jumping strategy goes as follows: Start at the vertex , then jump to the next vertex . The vertices and form a coset of (Note, each coset in consists of two elements, (mod ).) After we then jump to , which is in another coset of . After we then jump to the vertex . The vertices and form another coset of .
We repeat this process until each vertex is landed exactly once. This can be done since . So with . Therefore, is a permutation on
If , by Lemma 2, if ; otherwise . If , then the order of each coset in is . Also, with covers one coset of . So, . Similarly, if , by 1, with covers two cosets in . Thus, .
After making jumps, we jump to the next vertex, which is in . Then we continue jumping , which will cover one or two new cosets, depending on whether , as shown above. We continue this process until all vertices are covered exactly once. Thus the jump sequence determines a permutation.
Next we define the labeling sequence as follows:
To complete the proof for Case 1, it suffices to show:
Claim: is a valid radio--labeling with the desired span.
Proof of Claim) By Proposition 4, we need to show that and We call these the first and the second inequality, respectively. We begin by verifying the first inequality. Suppose is even. Then and
Suppose is odd. Then and
Now we verify the second inequality. By the facts that and , we obtain
Therefore, is a radio--labeling with the desired span.
Case 2. Both and are odd. Let and for , or and for . Then . Let
Case 2.1. Suppose is even. Define the jump sequence and labeling sequence respectively by:
We claim that our jump sequence gives a permutation. Let . If gcd, then by the second formula of in the above, gives a permutation. If gcd, since , the jump sequence labels each coset of , and then transitions into the next unlabeled coset by jumping (the first formula in above). Continue this process until all vertices are labeled. Hence, the jump sequence is a permutation.
It remains to show that is a valid radio--labeling. Consider the case that and . Then . By the assumption that is even, and must have the same parity, and
To verify the first inequality, as , we have
To verify the second inequality, using the facts that and , we obtain
Therefore, is a radio--labeling with the desired span. The case when follows similarly.
Case 2.2. Suppose is odd. If , , so and must have different parities. It follows that is odd. Otherwise, and . Then and have the same parity and again is odd. Hence always holds. Let .
Case 2.2.1. Suppose Define the jump sequence and the labeling sequence respectively by:
To show that the jump sequence is a permutation, we need to verify that . Since , we have that . This implies that . Hence and so , implying the jump sequence is a permutation.
Now we show is a valid radio--labeling. We prove the case when . (The case when follows similarly.) Then . For the first inequality, one can see that . For the second inequality, since and , one can verify that .
Therefore, we have a valid radio--labeling with the desired span, which is exactly the lower bound in Proposition 3.
Case 2.2.2. Suppose for some odd . Then , and there are cosets in . Recall that
Define the jumping sequence and labeling sequence respectively by:
Consider the case that and . (The case for and follows similarly.) Note that , , and .
Claim: Our jumping sequence is a permutation.
Proof of Claim) We show that we label two cosets of at a time by alternately jumping and (see the first and third formulas in above). That is, .
Note that . This implies . Hence, it is enough to show that . Because , we have . If , then , contradicting to gcd. Thus, , and our jump sequence begins by labeling two cosets in .
Next we show that the first two cosets labeled are and , by proving that . Let for some odd . Then
Thus, , and so . Hence, the first two labeled cosets are and .
Afterwards, we jump (the second formula in given above) and land in the coset in . Continue the same process repeatedly as above, we label the first cosets in pairs in the following order:
After the last pair of cosets above, the only coset that was left unlabelled is . Again, we jump and land in that coset.
It remains to show that jumping (the last formula in the given above) is a permutation of . This is true because is odd, so .
Finally we show that is a radio--labeling. The first inequality can be obtained directly from our definitions that for all , Next we verify the second inequality by listing all possible cases in Table 1. Note, for the last case (last row), we need to show . Recall . Then , which implies . For an example of this case, see Example 3.
Therefore, we have a valid radio--labeling with the desired span. The proof is complete. ∎
is odd | is even |
Example 2.
In Figure 1 we let and , so and . We get , , and . In this example we label one coset of at a time. With , covers one coset (circular black vertices) of . We then jump into a new coset (gray diamond vertices) of to complete the labeling.


Example 3.
In Figure 2 we let and . Then and . Notice that and . We label two cosets of at a time. With , covers two cosets (solid black and white circles) of . We then jump into a new coset (gray diamonds) of . We then jump consecutively to finish labeling the last coset.

4 Results for and (mod )
The case when and have different parities is more complicated, and so more interesting! Following the work in the previous section when and have the same parity, we will use the lower bound proved in 3. On the other hand, there are properties we use in the previous section that are not valid anymore for the case when and have different parities. The following lemma reveals the fact that in searching for an optimal radio--labeling that achieves the lower bound for we would lose some flexibility we had in 7. Consequently, we show that the lower bound in 3 is not always achieved.
Lemma 9.
Let be a radio--labeling for , where and have different parities. For any , if , then .
Proof.
9 shows that when and have different parities, in order to keep the -function value for any consecutive three vertices, , the distance between and must be fixed as . Throughout the section we denote this fixed value by . That is,
(4) |
This observation gives us the following lower bound. Recall the definition of from Eq. 2.
Proposition 10.
Suppose and , and let , where is defined in Eq. 4. Then
(5) |
Proof.
When is even and is odd, we establish lower and upper bounds as follows and show the bounds are sharp for some and .
Theorem 11.
Suppose is even and is odd with . Let . Then
Moreover, if then the lower bound in the above is sharp.
Proof.
The lower bound follows from Proposition 10. As , to prove the upper bound and the “moreover” part, by Lemma 2, it suffices to define a radio--labeling which satisfies all the following:
-
•
If , then labels one coset of at a time and increases the span by 1 each time it switches to an unlabeled coset.
-
•
If , then labels two cosets of at a time and increases the span by 1 each time it switches to a new pair of cosets.
Let . As is even and is odd, we let and ; and in either case, let . Define the jumping sequence and labeling sequence (for both cases) respectively by:
Then defines a radio--labeling; the proof follows similarly to the proofs in Theorem 8. Whenever , we have . This happens times while labeling. Thus, span.
If , then , so the upper bound holds. If , then , or ; hence the lower bound is sharp. ∎
Corollary 12.
Suppose is even and is odd with . Recall that . If , then .
Proof.
We conjecture that the upper bound in Theorem 11 is tight when (see Conjecture 1).
Now we turn to the case that is odd and is even. The remaining two results of this section are devoted to this case. Recall , , and defined in Eq. 1, Eq. 2, and Eq. 4, respectively.
Theorem 13.
Let be odd and even. If and is odd, then
Proof.
We define a jump sequence and labeling as follows, for all :
It is easy to see that our jumping sequence gives a permutation, since , thus
Theorem 14.
Suppose and where and . If , then
Proof.
To show that , we define a jumping and a labeling sequence in the following:
Finishing the proof by showing our jump sequence is a permutation and our labeling sequence is valid follows similarly to the proof of Theorem 8. ∎
Below is an example of Theorem 14.
Example 4.
In Figure 3 we let and . Then and . Notice that and . We label two cosets of at a time. With , covers two cosets (solid black circles and gray diamonds) of . We then jump into a new coset (larger gray circles) of . Notice that the difference of .

5 Closing Remarks
Most results in this paper are extensions of results by Karst, Langowitz, Oehrlein, and Troxell [10], and by Liu and Zhu [15]. The method of using the -function to prove a lower bound for the radio -number for was introduced in [15] when , the diameter of . Precisely, it was shown that for all . This method was applied and extended in [10], where the authors completely determined the radio -number for when , the diameter of plus one. The lower bound is shown again to be close to the exact value.
Theorem 15.
[10] Let with and For ,
Extending this lower bound approach with properties of cyclic groups, we were able to determine the values of for most . Many results in [15, 10] can be immediately obtained by this paper. For instance, when and have different parities we have with , or with . When and we have . Thus, Using Corollary 12, we have . When and we have . Thus . If is odd, by Proposition 13, .
For the case that , when and have different parities we have with , or with . When and , we have . Consider the case when is even. Then, . Using Corollary 12, we have . When and we have . Thus, if is not a multiple of . If is odd, by Proposition 13 the radio -number is
Although the values of for most and with have been determined, it remains open for other cases, especially when and have different parities. So far the evidence we obtained indicates that the upper bound in Theorem 11 might be always sharp for .
Conjecture 1.
Suppose is even and is odd with . Let and . If , then
Acknowledgement. The authors are grateful to the two anonymous referees for their careful reading of the manuscript and for their constructive suggestions which led to a better presentation of the article.
References
- [1] Devsi Bantva and Daphne Der-Fen Liu. Optimal radio labelings for block graphs and line graphs of trees. J. Theor. Comput. Sci., 891:90–104, 2021.
- [2] Devsi Bantva, Samir Vaidya, and Sanming Zhou. Radio number of trees. Discrete Appl. Math., 217(2):110–122, 2017.
- [3] Tiziana Calamoneri. The L(h, k)-Labelling Problem: A Survey and Annotated Bibliography. The Computer Journal, 49(5):585–608, 05 2006.
- [4] Gary Chartrand, David Erwin, and Ping Zhang. A graph labeling problem suggested by FM channel restrictions. Bull. Inst. Combin. Appl., 43:43–57, 2005.
- [5] Angel Chavez, Daphne Der-Fen Liu, and Mason Shurman. Optimal radio--radio labeling for trees. Euro. J. Combin., 91:13pp, 2021.
- [6] Jerrold Griggs and Roger Yeh. Labelling graphs with a condition at distance 2. SIAM Journal on Discrete Mathematics, 5(4):586–595, 1992.
- [7] Veronika Halász and Zsolt Tuza. Distance-constrained labeling of complete trees. Discrete Math., 338(8):1398–1406, 2015.
- [8] W. K. Hale. Frequency assignment: Theory and applications. Proceedings of the IEEE, 68(12):1497–1514, Dec 1980.
- [9] Justie Su-Tzu Juan and Daphne Der-Fen Liu. Antipodal labelings for cycles. Ars Combin., 103:81–96, 2012.
- [10] Nathaniel Karst, Joshua Langowitz, Jessica Oehrlein, and Denise Sakai Troxell. Radio -chromatic number of cycles for large . Discrete Math. Algorithms Appl., 9(3):1750031, 20, 2017.
- [11] Riadh Khennoufa and Olivier Togni. The radio antipodal and radio numbers of the hypercube. Ars Combin., 102:447–461, 2011.
- [12] Xiangwen Li, Vicky Mak, and Sanming Zhou. Optimal radio labellings of complete -ary trees. Discrete Appl. Math., 158(5):507–515, 2010.
- [13] Daphne Der-Fen Liu. Radio number for trees. Discrete Math., 308(7):1153–1164, 2008.
- [14] Daphne Der-Fen Liu, Laxman Saha, and Satyabrata Das. Improved lower bound for the radio number of trees. J. Theor. Comput. Sci., 851:1–13, 2021.
- [15] Daphne Der-Fen Liu and Xuding Zhu. Multilevel distance labelings for paths and cycles. SIAM J. Discrete Math., 19(3):610–621, 2005.
- [16] Paul Martinez, Juan Ortiz, Maggy Tomova, and Cindy Wyels. Radio numbers for generalized prism graphs. Discuss. Math. Graph Theory, 31(1):45–62, 2011.
- [17] Marc Morris-Rivera, Maggy Tomova, Cindy Wyels, and Aaron Yeager. The radio number of . Ars Combin., 120:7–21, 2015.
- [18] Amanda Niedzialomski. Radio graceful Hamming graphs. Discuss. Math. Graph Theory, 36(4):1007–1020, 2016.
- [19] Laxman Saha and Pratima Panigrahi. On the radio number of toroidal grids. Australas. J. Combin., 55:273–288, 2013.
- [20] Sanming Zhou. A channel assignment problem for optical networks modelled by Cayley graphs. Theoret. Comput. Sci., 310(1-3):501–511, 2004.
- [21] Sanming Zhou. Labelling Cayley graphs on abelian groups. SIAM J. Discrete Math., 19(4):985–1003, 2005.
- [22] Sanming Zhou. A distance-labelling problem for hypercubes. Discrete Appl. Math., 156(15):2846–2854, 2008.