Scientific African 24 (2024) e02235

Contents lists available at ScienceDirect

Scientific African
journal homepage: www.elsevier.com/locate/sciaf

Algorithms for computing chromatic polynomials and chromatic

index polynomials
Lateram Zawuga Hordofa , V.N. SrinivasaRao Repalle *, Mamo Abebe Ashebo
Department of Mathematics, Wollega University, Post Box No:395, Nekemte, Ethiopia

A R T I C L E I N F O A B S T R A C T

Editor: DR B Gyampoh Objectives: The aim of this article is to enhance the understanding of the computation of chromatic
polynomials and chromatic index polynomials, and to facilitate their practical use in various
Keywords: fields by demonstrating and supporting the proposed algorithms.
Graph algorithms Results: The theorems for computing the chromatic polynomial and chromatic index polynomial
Chromatic index polynomial
are thoroughly demonstrated and validated in this article. Additionally, the algorithms used in
Chromatic polynomial
these computations are introduced and explained.

Introduction

In graph theory, graph coloring is a subset of graph labeling that involves giving labels, known as colors, to nodes or edges of a
graph depending on certain criteria [8]. Graph coloring can be categorized into three types: vertex coloring, edge coloring, and map
coloring. Vertex coloring is a way of coloring the vertices of a graph so that no two neighboring vertices are the same color. Initially, the
notion of coloring involved coloring a map of the world so that each pair of countries sharing a border was assigned a distinct color
[22]. A planar graph is formed by connecting each pair of vertices representing countries sharing a boundary with an edge and
representing the countries on the map graph as vertices. The aim of graph coloring is to determine if every planar graph can be colored
with four colors [7]. The problem was later known as the four-color conjecture, and it was resolved in 1977 with the aid of computer
[3]. The issue of the number of ways of coloring was first explored by George David Birkhoff [5]. These counting the alternate colorings
of a map were later named the chromatic polynomial [23]. Vertex coloring problem can be seen as a conflict management issue in that
adjacent vertices in a graph must always have different colors, meaning they are in a constant state of conflict. More precisely, it can be
described as the problem of partitioning sets with internal irreconcilable conflicts [22]. In graph theory, edge coloring involves
assigning colors to the edges of a given graph so that adjacent edges have different colors [12]. This concept can be utilized in various
scheduling scenarios, notably in the timetable scheduling problem, aimed at organizing courses to prevent class overlaps [20]. By
associating each edge color with a distinct time slot, an edge coloring of a graph can effectively represent a viable course timetable
[14]. Moreover, utilizing edge colorings of complete graphs facilitates efficient scheduling of round-robin tournaments with minimal
rounds, ensuring all competitors face each other in a single round [11,17]. Additionally, edge coloring plays a pivotal role in
scheduling production processes [13] and in allocating channels or frequencies to transceiver links to mitigate primary and secondary
interference concerns [4].
The chromatic polynomials of various graph families have been computed by researchers [10,15]. Mamo and SrinivasaRao
introduced the chromatic polynomial of fuzzy graphs with crisp and fuzzy vertices using α-cuts [1], while Repalle et al. explored the

* Corresponding author.
E-mail address: rvnrepalle@gmail.com (V.N.S. Repalle).

https://doi.org/10.1016/j.sciaf.2024.e02235
Received 30 December 2023; Received in revised form 21 March 2024; Accepted 2 May 2024
Available online 3 May 2024
2468-2276/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

computation of chromatic polynomials for intuitionistic fuzzy graphs and their properties [16]. Hordofa et al. presented the strong
fuzzy chromatic polynomial for intuitionistic fuzzy graphs [9]. Sanli and Cangul proposed methods for calculating chromatic poly­
nomials of graphs, including disjoint union, a graph with one pendant vertex, a graph with multiple pendant vertices, and the union of
two graphs connected by a bridge [19]. Whitney’s fundamental reduction formula has remained a valuable tool for solving chromatic
polynomials [23]. Despite the existing literature on techniques for computing graph chromatic polynomials, researchers are still in
search of additional algorithms.
The chromatic index polynomial illustrates different ways to color the edges of a loop-less graph with a specified number of colors
or fewer [18]. It provides insights into edge coloring alternatives. This polynomial has been calculated for various graph types such as
star graphs, bi-star graphs, fan graphs, Helm graphs, and friendship graphs [18]. However, the chromatic index polynomial for most
graphs remains unexplored. Based on my investigation, there is a lack of published literature outlining the algorithms for determining
the chromatic index polynomial of a graph. This spurred our efforts to develop appropriate algorithms for computing chromatic index
polynomials.
The article is structured into four sections. The first section introduces graph colorings, chromatic polynomials, and chromatic
index polynomials. The second section defines key terms and propositions. The third section explores chromatic polynomial algorithms
and chromatic index algorithms, along with proofs of the related theorems. Finally, the fourth section provides a summary of the
article.

Preliminaries

This section presents crucial definitions and propositions that are essential for comprehending the article.
Definition 1. [6] Vertex coloring is the task of assigning different colors to adjacent vertices, whereas edge coloring is the task of
assigning distinct colors to edges that share a common vertex.

Definition 2. [6] The chromatic polynomial of a simple graph G represents the number of unique ways to properly color its vertices
using a maximum of k colors. It is written as P(G, k).

Definition 3. [18] The chromatic index polynomial of graph G, denoted by P(G,x), represents the count of distinct ways to achieve a
proper edge coloring of G using x or fewer colors.

Definition 4. [21] A matching M in graph G is a set of pairwise non neighbor edges of E(G).

Proposition 1. [23] (deletion-contaction method): Consider a graph, denoted as G, where u and v are two non-adjacent vertices in G. Now,
let Gʹ be the graph obtained by deleting an edge between u and v, and Gʹ be the graph obtained by contracting u and v into one vertex. Then, the
chromatic polynomial, P(G, k) = P(Gʹ, k) − P(Gʹ , k).

Proposition 2. [2] Suppose G0 is a graph with chromatic polynomial P(G0 , x). If G1 is a graph created by adding a new vertex v to G0 and
connecting v to every vertex of a complete subgraph of G0 containing n vertices, then the chromatic polynomial of G1 is given by P(G1 ,x) = (x −
n)P(G0 , x).

Proposition 3. [2] Let G = (V, E) be a simple graph with |V| = n and contains a subgraph Km . If each of the mutually disjoint vertex in
V(G − Km ) is joined with every vertex in Km , then P(G, x) = (x − m)n− m P(Km , x).

Main results

Chromatic polynomial

In this section, we demonstrate theorems and introduce specific algorithms related to the computation of chromatic polynomials.
Theorem 1. Let G = (V, E) be a simple graph with |V| = n and |E| = m. If Kn− 1 is a subgraph of G, then P(G, x) = P(Kn , x) +
( )
n2 − n− 2m
2 P(Kn− 1 , x).

Proof. Consider a simple graph G = (V, E) with |V| = n and |E| = m. Let v be in V(G) − V(Kn− 1 ) and ui v be in E(G∁ ) for ui in V(Kn− 1 )
and for some i = 1, 2, …, n − 1. Then, G∁ is equal to Kn − G and it implies that |E(G∁ )| = |E(Kn )| − |E(G)| = n(n−2 1) − m. Since Kn− 1 is a
clique of G, for each i G \ ui v is reduced to Kn− 1 . Now after applying the addition-contraction algorithm as many as the number of edges
( )
in (G∁ ), P(G, x) = P(Kn , x) + n(n−2 1) − m P(Kn− 1 , x).

Example 1. Consider a simple graph G given as shown in Fig. 1. G contains 4 vertices and 4 edges. Also it has a sub graph K3 which is

2
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

a clique on 3 vertices. Since G∁ has n(n−2 1) − m = 12 (4)(3) − 4 = 2 edges and applying the Theorem 1, P(G, x) = P(K4 ,x) + (2)P(K3 ,x) =
x(x − 1)(x − 2)((x − 3) + 2) = x(x − 1)2 (x − 2).

The chromatic polynomial algorithm for a graph on n vertices containing a clique Kn− 1

1. Select a vertex v ∈ V with a minimum vertex degree.

2. For a sub graph on the set of vertices V − {v} and edge e that connects non adjacent vertices in V − {v} with vetex v, carry out
P(G, x) = P(G + e, x) + P(G \ e, x)

( )
n2 − n− 2m
3. Repeat 1 and 2 until G + e and G \ e become complete graphs. Then, calculate P(G, x) = P(Kn , x) + 2 P(Kn− 1 , x). Where
n2 − n− 2m
2 is the number of edge e that connects non adjacent vertices in V − {v} with vetex v.

Example 2. Let us illustrate the algorithm above on a graph G in Fig. 2.

By Using Figure2; P(G, x) = P(K5 , x) + 3P(K4 , x) + 2[P(K4 , x) + 2P(K3 , x)] + P(K3 , x) + P(K2 , x)

= P(K5 , x) + 5P(K4 , x) + 5P(K3 , x)] + P(K2 , x)

( )
= x(x − 1)2 x2 − 3x + 3

Theorem 2. Let G = (V, E) be a simple graph on V = {u, v, v1 , v2 , …, vn− 2 } containing |E| = m and a sub graph Kn− 2 on {v1 ,v2 ,…,vn− 2 }. If
( ) ( )
1) (n− 1)(n− 2)
deg(v) = N is the smallest and v is not joined with u, then P(G, x) = P(Kn ,x) + (n)(n−
2 − m P(K n− 1 ,x) + 2 − m + N + M (n −
N − 2)P(Kn− 2 , x).
Proof: Consider a simple graph, G = (V, E) having |V| = n and |E| = m. Assume that Kn− 2 is a sub graph of G and there is a vertex
v ∈ V having a minimum vertex degree equal to N. Now we find G − v by removing all edges that are incident on v. Since Kn− 2 is a
( ) ( )
subgraph of G, G − v needs (n− 1)(n−
2
2)
− |E(G − v)| = (n− 1)(n− 2)
2 − m +N more edges to be Kn− 1 . Now after adding and contracting all
( )
these (n− 1)(n−
2
2)
− m +N edges on G, we obtain a new graphs, say H1 and H2 that contain sub graphs Kn− 1 and Kn− 2 . Then again
applying addition-contraction method,
( )
(n − 1)(n − 2)
P(G, x) = P(H1 , x) + − m + N P(H2 , x) (a)
2

1) (n− 1)(n− 2) 1)(n− 2)

Now applying Theorem 1, we can verify that H1 needs more (n)(n−
2 − 2 − N edges to be Kn and H2 needs more (n− 2 −
(n− 2)(n− 3)
2 − N edges to be Kn− 1 . Now after adding all these edges to these graphs and contracting them,
( ) ( )
1) (n− 1)(n− 2) (n− 1)(n− 2) 2)(n− 3)
P(H1 , x) = P(Kn , x) + (n)(n−
2 − 2 − N P(K n− 1 , x) and P(H2 , x) = P(K n− 1 , x) + 2 − (n− 2 − N P(Kn− 2 , x). By

Fig. 1. A graph on 4-vertices and containing clique K3.

3
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

Fig. 2. Illustration of the algorithm for a graph on n vertices with a clique Kn− 1 on a graph in Fig. 1.

substituting these values of P(H1 , x) and P(H2 , x) in equation (a), we obtain

( )
(n)(n − 1) (n − 1)(n − 2)
P(G, x) = P(Kn , x) + − − N P(Kn− 1 , x)
2 2
( )( ( ) )
(n − 1)(n − 2) (n − 1)(n − 2) (n − 2)(n − 3)
+ − m + N P(Kn− 1 , x) + − − N P(Kn− 2 , x) (b)
2 2 2
( ) ( )
1) (n− 1)(n− 2)
Up on simplifying equation (b), we prove that P(G, x) = P(Kn , x) + (n)(n− 2
− m P(K n− 1 , x) + 2
− m + N (n − N −
2)P(Kn− 2 , x).

The chromatic polynomial algorithm for a graph on n vertices containing a clique Kn− 2
Consider a simple graph, G = (V, E) with |V| = n and |E| = m containing Kn− 2 as a sub graph.

1. Choose a vertex v ∈ V having a minimum vertex degree equal to N.

(n− 1)(n− 2)
2. Identify all the 2 − m + N edges, G − v needs to become Kn− 1 .
( )
(n− 1)(n− 2)
3. Add and contract all the edges in 2, to obtain P(G, x) = P(H1 , x) + 2 − m+ N P(H2 , x), where
( ) ( )
1) (n− 1)(n− 2)
P(H1 , x) = P(Kn , x) + (n)(n−
2 − 2 − N P(Kn− 1 , x) and P(H2 , x) = P(Kn− 1 , x) + (n− 1)(n−
2
2)
− (n− 2)(n− 3)
2 − N P(Kn− 2 , x).

4
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

( ) ( )
(n)(n− 1) (n− 1)(n− 2)
4. Lastly from 3, P(G, x) = P(Kn , x) + 2 − m P(Kn− 1 , x) + 2 − m+N (n − N − 2)P(Kn− 2 , x).

Example 3. Consider a graph G containing 5 vertices and the clique sub graph K3
(
1)
Since in the graph G shown in Fig. 3, Deg(5) = 1 = N, |E| = 6 and |V| = 5, P(G, x) can be computed as; P(K5 , x) + (5)(5− 2 −
) ( )
6 P(K5− 1 ,x) + (5− 1)(5−
2
2)
− 6 + 1 (5 − 1 − 2)P(K5− 2 ,x). Thus, P(G, x) = P(K5 ,x) + (4)P(K4 ,x) + (2)P(K3 ,x) = x(x − 1)(x − 2)(x −
3)(x − 4) + 4x(x − 1)(x − 2)(x − 3) + 2x(x − 1)(x − 2).

Chromatic index polynomial

In this section, we present the demonstration and proofs of some theorems, as well as we introduce specific algorithms related to the
computation of chromatic index polynomials.

The computation of chromatic index polynomial using matching sets

Theorem 3. Let G be a graph with m > 0 edges. If x colors are given, then the chromatic index polynomial of G, is denoted by P(G, x) and
( )
∑ x
defined byP(G, x) = m j=1 β(G, j) j . Where β(G, j) is distinct ways of properly coloring edges of G using exactly j distinct colors.
( )
x
Proof: Consider a graph G containing m > 0 number of edges. Let x colors be given, then we choose j colors out of x colors, in
j
distinct ways. Now let β(G, j) be the different ways of properly coloring edges of G using exactly j different colors. In other words,
β(G, j) is the number of distinct ways of partitioning E(G) into j matching. Then we color G using exactly j colors out of x colors in
( )
x
β(G, j) distinct ways. Since j can be any positive integer from 1 to m, the chromatic index polynomial of G with x colors or fewer
j
( ) ( )
x ∑ x
is the sum of β(G, j) over j. Thus, P(G, x) = m β(G, j) .
j j=1 j

Definition 6. Let G = (V, E) be a graph containing|E(G)| = m. Then,

( )
∑ x
P(G, x) = m β(G, j) , where β(G, j) is the number of distinct ways of partitioning E(G) into j matching.
j=1 j

Example 4. Consider a graph G shown in Fig. 4, to illustrate the calculation of P(G, x) applying Theorem 3.
To illustrate Theorem 3, let us determine the partition E(G) into j matching for j = 1,2,3,4,5,6. It is impossible to partition E(G) into
either one or two matching. Thus, β(G, j) = 0, for j = 1, 2.
There is only one way of partitioning E(G) into three matching. That is, {a, c}, {b, d}, {e, f}. So, β(G, 3) = 1.
There are three ways of partitioning E(G) into four matching. Namely,

i. {a}, {c}, {b, d}, {e, f}

ii. {b}, {d}, {a, c}, {e, f}
iii. {e}, {f}, {a, c}, {b, d}

Fig. 3. A graph G containing 5 vertices and the clique sub graph K3.

5
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

Fig. 4. A graph G utilized in the computation of P(G, x) by partitioning of edges of G into matching.

So, β(G, 4) = 3.
There are three ways of partitioning E(G) into five matching. These are;

i. {a}, {c}, {b}, {d}, {e, f}

ii. {a}, {c}, {e}, {f}, {b, d}
iii. {b}, {d}, {e}, {f}, {a, c}

So, β(G, 5) = 3.
Lastly, there is one way of partitioning E(G) into six matching. That is, {a}, {c}, {b}, {d}, {e}, {f} and it shows β(G, 6) = 1.
( )
∑ x
Now, P(G, x) = m j=1 β(G, j) j

6 ( )
∑ x
= β(G, j)
j=1
j

( ) ( ) ( ) ( ) ( ) ( )
x x x x x x
= β(G, 1) + β(G, 2) + β(G, 3) + β(G, 4) + β(G, 5) + β(G, 6)
1 2 3 4 5 6
( ) ( ) ( ) ( )
x x x x
=0+0+ +3 +3 +
3 4 5 6

= (x)(x − 1)(x − 2) + 3(x)(x − 1)(x − 2)(x − 3) + 3(x)(x − 1)(x − 2)(x − 3)(x − 4) + (x)(x − 1)(x − 2)(x − 3)(x − 4)(x − 5)
( )
= x(x − 1)(x − 2) x3 − 9x2 + 29x − 32

The computation of chromatic index polynomial using a recursive edge deletion algorithm
Definition 5. The edge neighborhood degree of an edge e = uv in graph G is denoted by |NG (e)|, which is computed by adding the
count of edges connected to vertex u with the count of edges connected to vertex v.

Definition 6. The maximum edge neighborhood degree of a graph G is denoted by σʹ(G) is defined as the maximum edge neigh­
borhood degree of edges of G.

Theorem 4. Let G be a graph containing m (m ∕ = 0) edges and that each of these edges is a neighbor. Then, P(G, x) = x{m− 1} .
Proof: Let’s assume a graph G containing m (m ∕ = 0) edges and having each of its edges is a neighbor. That means, |NG (e)| = m − 1,
∀e ∈ E(G). Since every edge of G is a neighbor, it implies that we have to color each edge distinctly. This in turn, we color one of the
edges in x ways, the other in x − 1 ways, and so on, and similarly the last edge in x − m + 1 ways. Therefore, the chromatic index
polynomial of G is the product of the ways of coloring each edge. Hence, P(G, x) = x(x − 1)…(x − m + 1).

Theorem 5. Let G be a graph containing an edge e such that |NG (e)| =σʹ(G) − 1. Then there is another edge f which is not in NG (e) and P(G,
( )
x) = x P(G − {e}, x − 1) + P(G − {e, f}, x − 1) .
Proof: The chromatic index polynomial of G is the sum of the chromatic index polynomial of G when e and f are colored with similar
colors and the chromatic index polynomial of G when e and f are colored differently. In other words, the chromatic index polynomial of

6
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

G is the sum of coloring every edge of G distinctly and coloring every edge of G containing one less edge distinctly. Thus, P(G, x) =
ʹ ʹ ( ʹ ʹ ) ( )
x{σ (G)+1} + x{σ (G)} = x x{σ (G)} + x{σ (G)− 1} = x P(G \ {e}, x − 1) + P(G \ {e, f}, x − 1) .

Theorem 6. Let G be a graph containing an edge e such that |NG (e)| =σʹ(G) − 2. Then there are edges fand g which are not in NG (e) and the
chromatic index polynomial of G is equal to:
( )
P(G, x) = x P(G \ {e}, x − 1) + P(G \ {e, f}, x − 1) + P(G − {e, g}, x − 1) , when f and g are not in the same matching.
( )
P(G,x) = x P(G \ {e}, x − 1) + P(G \ {e,f}, x − 1) + P(G − {e,g}, x − 1) + P(G − {e,f,g}, x − 1) , when f and g are in the same
matching.
Proof. Suppose there are edges f and g which are not neighbor with e. Then there are two cases when f and g are neighbor and f and g
are not neighbor (or f and g are in the same matching).
Case 1. Consider when f and g are not in the same matching. Then the corresponding matching involving e would be {e, f} and {e,g}.
Then the chromatic index polynomial of G is the sum of chromatic index polynomial of G when e and f colored with similiar color and
chromatic index polynomial of G when e, f and g are colored differently, and e and g colored with similiar color. This implies,
( )
P(G, x) = x P(G \ {e}, x − 1) + P(G \ {e, f}, x − 1) + P(G − {e, g}, x − 1) , since f and g are not in the same matching.
Case 2. Let’s consider the scenario where both f and g are in the same matching. In this case, the possible matchings involving e
would be {e,f}, {e,g}, and {e,f,g}. Since the edges within the same matching are not neighbor to each other, they can be colored with
the same color. However, it is also possible to color them distinctly if desired. This show,
( )
P(G, x) = x P(G \ {e}, x − 1) + P(G \ {e, f}, x − 1) + P(G − {e, g}, x − 1) + P(G − {e, f, g}, x − 1)

Theorem 7. Let G be a graph containing an edge e such that |NG (e)| =σʹ(G) = m − l. Assume that there is a set e which contains a set of all
( )
∑
edges which are not neighbor to e and that F are matching subsets of e. Then P(G,x) = x P(G − {e}, x − 1) + P(G − {{e} ∪ F}, x − 1)
F

Proof. Suppose we have a graph, G, with an edge, e, such that |NG (e)| = σʹ(G) = m − l. Let e = f1 , f2 , …, fl be a set that contains all
{ }

edges in G that are not neighbors of e. Using Theorems 5 and 6, we demonstrate that the theorem is valid for l = 1 and l = 2. The

Fig. 5. Utilization of the recursive edge deletion algorithm.

7
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

theorem is assumed to hold for l = k + 1, and we check to see if it follows from l = k. Now let F be a set containing all matching subsets
of f1 , f2 , …, fk , fk+1 and let Fʹ be a set containing all matching subsets of f1 , f2 , …, fk that do not include {fk+1 }. In other words, Fʹ is a
{ } { }

set containing all matching subsets of f1 , f2 , …, fk . It is clear that Fʹ is a subset of F. When l = k, P(G, x) has the terms
{ }
( )
x P(G − {e}, x − 1) + P(G − {{e} ∪ Fʹ}, x − 1) and
∑
Fʹ
( )
When l = k + 1, P(G, x) = x P(G − {e}, x − 1) + P(G − {{e} ∪ Fʹ}, x − 1) + P(G − {{e} ∪ (F − Fʹ)}, x − 1) .
∑ ∑
Fʹ F− Fʹ

Every element in F − Fʹ has a matching that includes fk+1 . Hence, the addition of edge fk+1 results in the terms
P(G − {{e} ∪ (F − Fʹ)}, x − 1) being added. Thus, the theorem proved.
∑
F− Fʹ

Recursive edge deletion algorithm

Let a graph G containing a nonempty edge set be given. Then the chromatic index polynomial of G can be computed by applying the
recursive edge deletion algorithm as:
Step 1: Select an edge from G that has the highest number of neighbor edges. That is; choose e1 ∈ E(G) such that |NG (e)| = σʹ(G).
{ ⃒
Step 2: Create a set of edges, e1 = fi ⃒fi ∈ E(G) and fi ∕
∈ NG (e1 )}. Based on this set find all the matching subsets of e1 denoted by
Ej .
{ }
Step 3: For each Ej , find the sub graphs G − {e1 } and G − {e1 } ∪ Ej
( { } )
Step 4: Calculate P(G − {e1 }, x − 1) and P G − {e1 } ∪ Ej , x − 1 .
Step 5: Calculate P(G, x) as:
( )
∑ ( { } )
P(G, x) = x P(G − {e1 }, x − 1) + P G − {e1 } ∪ Ej , x − 1 , where Ej represents each matching subset of e1 .
Ej

Example 4. Consider a graph H in Fig. 5. Now we are interested in evaluating the chromatic index polynomial of H using the
recursive edge deletion method. The graph is represented in Fig. 5, and the neighborhood degrees of the edges of H are shown in
Table 1.
To calculate the chromatic index of H using the recursive edge deletion algorithm, we can use the following steps: Start with the
original graph H. Choose the edge with the highest neighborhood degree, labeled as 1 and 2, as both |NH (1)| and |NH (2)| equal to 4,
which is the same as σʹ(H). Edges 4 and 5 are not neighbor edge of 2, and the matching subset consisting of edges 4 and 5 is represented
by {4} and {5}. Thus, following the algorithm, edge 2 is removed from H to obtain H − {2}, edge 2 and 4 are removed from H to obtain
H − {2, 4}, and edge 2 and 5 are removed from H to obtain H − {2, 5}, as illustrated in the Fig. 5.
[
Now using recursive edge deletion algorithm for a graph H above, the chromatic index of H is, P(H, x) = x P(H − {2}, x − 1) +
]
P(H − {2, 4}, x − 1) + P(H − {2, 5}, x − 1) . Where the each chromatic index polynomial is calculated as follows:
( )
P(H − {2}, x) = x(x − 1)(x − 2)2 x2 − 2x + 2

P(H − {2, 4}, x) = x(x − 1)4

( )
P(H − {2, 5}, x) = x(x − 1)(x − 2) x2 − 2x + 2
[ ( ) ( )]
Therefore, P(H, x) = x (x − 1)(x − 2)(x − 3)2 x2 − 4x + 5 + (x − 1)(x − 2)4 + (x − 1)(x − 2)(x − 3) x2 − 4x + 5

( )
Therefore, the chromatic index of graph H is P(H, x) = x(x − 1)(x − 2)2 x3 − 6x2 + 13x − 11 .

Table 1
The edge and its neighborhood degree of a graph in
Fig. 5.
Edge e in E(H) |NH (e)|

1 4
2 4
3 3
4 3
5 2
6 3
7 3

8
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

Calculation of the chromatic index polynomial using a generalized recursive edge deletion algorithm
Generalized recursive edge deletion algorithm
Consider a graph G containing a non-empty edge set. Then the steps in the generalized recursive edge deletion algorithm to
compute the chromatic index polynomial of G are listed below.

Step 1. Choose an edge in E(G), such that its edge neighborhood degree is the largest. That is: choose e1 if |NG (e1 )| = σʹ(G)
{ ⃒
Step 2. Determine a set e1 = fi ⃒fi ∈ E(G) and fi ∕ ∈ NG (e1 )}
{ ⃒
Step 3. Create a set of edges, e1 = fi ⃒fi ∈ E(G) and fi ∕
∈ NG (e1 )}. Based on these set find all the matching subsets of e1 denoted by Ej .
{ }
Step 4. For each Ej , find the subgraphs G − {e1 } and G − {e1 } ∪ Ej .
Step 5. For every sub graph in step 4, repeat steps 1–4 until the resulting sub graphs become either cycle graphs or path graphs
(possibly disconnected path graph components). We may have the following three cases:
i. Suppose F is a cycle graph, Cn obtained by iterating steps 1–3 k-times, Then, P(G) contains a term x(x − 1)…(x − k +
1)[(x − k − 1)n + (− 1)n (x − k − 1)]
[
ii. Suppose F is a Path graph, Pn obtained by iterating steps 1–3 k-times, Then, P(G) contains a term x(x − 1)…(x − k + 1) (x −
]
k)(x − k − 1)n− 2
iii. Suppose F is a disconnected m Path graph component, obtained by iterating steps 1–3 k-times, Then, P(G) contains a term x(x −
[ ]
1)…(x − k + 1) (x − k)m (x − k − 1)n− |E(G)|+1 .
∑
Step 6. Sum up the contributions from each sub-graph in step 5 to obtain the result of P(G). That is, P(G, x) = x(x − 1)…(x − k +
( )
1) P(F, x − k) , where F is a sub-graph in one of the three cases listed in step 4 and k is the number of times Steps 1–3 have
been repeated.

Fig. 6. The computation of the chromatic index polynomial by general recursive edge deletion algorithm.

9
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

In general, this approach is based on recursively breaking down the graph into smaller sub graphs and computing their contri­
butions to the final result and provides an efficient and systematic approach to determine the chromatic index polynomial of a graph
involving a non-empty edge set.
Example 5. let’s examine the graph G shown in Fig. 6 and apply the general recursive edge deletion technique to compute its
chromatic index polynomial.
The edges in Fig. 6 that are broken lines are those that are not neighbors to the heavy line. The first iteration’s sub graphs, which are
shown on the leftmost side, show how to remove the heavy line from the original graph G. The following 4 subgraphs show, on the
right, how the independent matchings with the broken edges are eliminated.
To better explain the method, the leftmost sub graphs are once again taken into account for the second iteration. In this iteration,
the edges that are selected and form an independent matching are deleted from the sub graph, along with the heavy-lined edge.
For each sub graph in the second iteration, we calculate their chromatic index polynomial at x − 2 and multiply the outcome by
x(x − 1). Similarly, for each subgraph in the first iteration, we calculate their chromatic index polynomial at x − 1 and multiply the
outcome by x. The chromatic index polynomial of G is then the sum of the results that are obtained from each subgraph.
For a given x colors, the chromatic index polynomial of a graph obtained by concatenating a cycle graph Cn with a path graphs Pm
[
with edge from one vertex of Ck to one pendant edge of Pl for integers n ≥ 3 is be calculated by (x − 2) (x − 1)n+m− 2 +
]
(− 1)n (x − 1)m− 1 . Therefore in the first iteration the chromatic index polynomial of the third sub graph put in the bracket is:
[ ] [ ]
(x − 2) (x − 1)5 − (x − 1) and evaluate it at x − 1. This results in (x − 3) (x − 2)5 − (x − 2) . Using the subgraphs from the second
iteration, the Chromatic index polynomial of the most left sub graph in the first iteration is calculated and evaluate the at x − 2 and each
will be multiplied by x(x − 1) .
[ ( ) ( ) ( ) ( )
P(G, x) = x(x − 1) P(P7 , x − 2) + P(P4 , x − 2) × P(P3 , x − 2) + P(P5 , x − 2) × P(P2 , x − 2) + P(P6 , x − 2)
( ) ( ) ( ) ( )]
+ P(P3 , x − 2) × P(P3 , x − 2) + P(P4 , x − 2) × P(P2 , x − 2)
[( ) ( ) [
+ x P(C5 , x − 1) × P(P2 , x − 1) + P(P7 , x − 1) + (x − 3) (x − 2)5 − (x − 2) + P(P6 , x − 1)].

Lastly, by substituting P(P7 , x) = x(x − 1)n− 2 and P(C5 , x) = (x − 1)n + (− 1)n (x − 1).
That is,
( ) ( )
P(G, x) = x(x − 1)(x − 2)2 (x − 3)2 x2 − 4x + 5 + x(x − 1)(x − 2)2 3x3 − 19x2 + 42x − 34
( )
= x(x − 1)2 (x − 2)2 x3 − 6x2 + 13x − 11 .

Conclusion and future works

In graph coloring, vertex coloring and edge coloring are essential parts that entail assigning colors to nodes or edges. The chromatic
polynomial counts all the distinct vertex colorings, while the chromatic index polynomial computes the number of distinct edge
colorings. These counting the number of colorings find applications in diverse fields like timetable scheduling, round-robin tourna­
ment scheduling, production process scheduling, puzzle solving, chemical storage for drug interactions, frequency assignment, and
register allocation. This article has explored the theorems related to computing both the chromatic polynomial and the chromatic
index polynomials, providing evidence and validation to support these theorems. However, the study lacks analysis of the efficiency of
the methods. Additionally, the article introduces and explains the algorithms utilized in the computations of chromatic polynomials
and chromatic index polynomials. But the study does not compare computation methods or algorithms with alternative approaches,
prevents a comprehensive evaluation of their advantages.
In the future, we will compare the algorithms with other parallel algorithms. Additionally, we will provide the computer code for
these algorithms.

Funding for manuscript

This work received no financial support.

CRediT authorship contribution statement

Lateram Zawuga Hordofa: Conceptualization, Formal analysis, Methodology, Writing – review & editing. V.N. SrinivasaRao
Repalle: Supervision, Writing – original draft, Writing – review & editing, Methodology. Mamo Abebe Ashebo: Supervision, Writing
– original draft, Writing – review & editing, Methodology.

10
L.Z. Hordofa et al. Scientific African 24 (2024) e02235

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

References

[1] M. Abebe Ashebo, V.N.S.R. Repalle, Fuzzy chromatic polynomial of fuzzy graphs with crisp and fuzzy vertices using α-cuts, Adv. Fuzzy Syst. (2019) 1–11,
https://doi.org/10.1155/2019/5213020.
[2] A.A. Adam, The Chromatic Polynomial of a Graph, University of The Witwatersrand, 1990 (Doctoral Dissertation.
[3] K. Appel, W. Haken, J. Koch, Every planar map is four colorable. Part II: reducibility, Ill. J. Math. 21 (3) (1977) 491–567, https://doi.org/10.1215/ijm/
1256049012, 1997.
[4] C.L. Barrett, et al., Strong edge coloring for channel assignment in wireless radio networks, in: Proceedings of the Fourth Annual IEEE International Conference
on Pervasive Computing and Communications Workshops, PerCom Workshops, 2006, pp. 106–110, 10.1109/PERCOMW.2006.129.
[5] G.D. Birkhoff, A determinant formula for the number of ways of coloring a map, Ann. Math. 14 (1/4) (1912) 42, https://doi.org/10.2307/1967597.
[6] J.A. Bondy, U.S.R. Murty, Graph Theory with Applications, Macmillan Education UK, 1976, 10.1007/978-1-349-03521-2.
[7] G. Chartrand, P. Zhang, Chromatic Graph Theory, Chapman and Hall/CRC, 2019, 10.1201/9780429438868.
[8] A. Elumalai, Graph coloring and its implementation, Malaya J. Mat. S (2) (2020) 1672–1674, https://doi.org/10.26637/MJM0S20/0445.
[9] L.Z. Hordofa, V.N.S.R. Repalle, M.A. Ashebo, Strong fuzzy chromatic polynomial of intuitionistic fuzzy graphs (IFGs) based on (α,β)-levels, Inf. Sci. Lett. 12 (6)
(2023) 2227–2233, https://doi.org/10.18576/isl/120603.
[10] B. Jackson, Chromatic polynomials. Topics in Chromatic Graph Theory, Cambridge University Press, 2015, pp. 56–72. https://doi.org/10.1017/
cbo9781139519793.006.
[11] T. Januario, S. Urrutia, C.C. Ribeiro, D. de Werra, Edge coloring: a natural model for sports scheduling, Eur. J. Oper. Res. 254 (1) (2016) 1–8, https://doi.org/
10.1016/j.ejor.2016.03.038.
[12] G. Chartrand, L. Lesniak, P. Zhang, Graphs & Digraphs, in: 6th ed., CRC Press, New York, 2015, pp. 465–492. https://doi.org/10.1201/b19731.
[13] C.C. Lin, D.J. Deng, Y.L. Chih, H.T. Chiu, Smart manufacturing scheduling with edge computing using multiclass deep Q network, IEEE Trans. Ind. Inform. 15
(7) (2019) 4276–4284, https://doi.org/10.1109/tii.2019.2908210.
[14] G.A. Neufeld, J. Tartar, Graph coloring conditions for the existence of solutions to the timetable problem, Commun. ACM 17 (8) (1974) 450–453, https://doi.
org/10.1145/361082.361092.
[15] R.C. Read, An introduction to chromatic polynomials, J. Comb. Theory 4 (1) (1968) 52–71, https://doi.org/10.1016/s0021-9800(68)80087-0.
[16] V.N.S.R. Repalle, L.Z. Hordofa, M.A. Ashebo, Chromatic polynomial of intuitionistic fuzzy graphs using (α,β)-levels, Int. J. Math. Math. Sci. 7 (2022), https://
doi.org/10.1155/2022/9320700. Article ID 9320700.
[17] C.C. Ribeiro, Sports scheduling: problems and applications, Int. Trans. Oper. Res. 19 (1–2) (2012) 201–226, https://doi.org/10.1111/j.1475-3995.2011.00819.
x. Portico.
[18] R.A. Robles, Chromatic index of some classes of graph, Int. J. Adv. Res. Publ. 3 (7) (2019) 78–82 (IJARP), http://www.ijarp.org/online-papers-publishing.
[19] U. Sanli, I.N. Cangul, A new method for calculating the chromatic polynomial, Appl. Sci. 19 (2017) 110–121. http://www.mathem.pub.ro/apps/v19/A19-sa-
C84P.pdf.
[20] N.D. Thanh, Solving timetabling problem using genetic and heuristic algorithms, in: Proceedings of the Eighth ACIS International Conference on Software
Engineering 3, Artificial Intelligence, Networking, and Parallel/Distributed Computing (SNPD 2007), 2007, pp. 472–477.
[21] K.R. Varadarajan, P.K. Agarwal, Approximation algorithms for bipartite and non-bipartite matching in the plane, SODA 99 (1999) 805–814. https://www2.cs.
arizona.edu/~alon/papers/approx-match.pdf.
[22] V.I. Voloshin, Graph coloring: history, results and open problems, Ala. J. Math. 34 (2009) 1–3. Spring/Fall.
[23] H. Whitney, The coloring of graphs, Ann. Math. 33 (4) (1932) 688–718, https://doi.org/10.2307/1968214.

11