M.Sc.

Thesis Mid Semester Presentation (Stage 1)

Complexity Analysis of a Graph Coloring Problem

Akash Tayal

Roll Number: 2303141002

Supervisor: Dr.Priyamvada
Department of Mathematics
IIT Indore

October 5, 2024

Akash Tayal Complexity Analysis October 5, 2024 1 / 29

Outline

1 Introduction

2 Preliminaries

3 Complexity of Graph Coloring Problem

4 Future Work

5 References

Akash Tayal Complexity Analysis October 5, 2024 2 / 29

Outline

1 Introduction

2 Preliminaries

3 Complexity of Graph Coloring Problem

4 Future Work

5 References

Akash Tayal Complexity Analysis October 5, 2024 3 / 29

Introduction

Graph Coloring Problem: The task of assigning colors to vertices of

a graph so that no two adjacent vertices share the same color.
It is a classic combinatorial problem with scheduling, networking, and
optimization applications.
Determining the minimum number of colors (chromatic number) for
a graph is NP-Hard , making it computationally challenging.
Our Focus is to analyze the complexity of related problems in specific
classes of graphs and explore whether it is NP-Complete.

Akash Tayal Complexity Analysis October 5, 2024 4 / 29

Introduction

Graph Coloring Problem: The task of assigning colors to vertices of

a graph so that no two adjacent vertices share the same color.
It is a classic combinatorial problem with scheduling, networking, and
optimization applications.
Determining the minimum number of colors (chromatic number) for
a graph is NP-Hard , making it computationally challenging.
Our Focus is to analyze the complexity of related problems in specific
classes of graphs and explore whether it is NP-Complete.

Akash Tayal Complexity Analysis October 5, 2024 4 / 29

Introduction

Graph Coloring Problem: The task of assigning colors to vertices of

a graph so that no two adjacent vertices share the same color.
It is a classic combinatorial problem with scheduling, networking, and
optimization applications.
Determining the minimum number of colors (chromatic number) for
a graph is NP-Hard , making it computationally challenging.
Our Focus is to analyze the complexity of related problems in specific
classes of graphs and explore whether it is NP-Complete.

Akash Tayal Complexity Analysis October 5, 2024 4 / 29

Introduction

Graph Coloring Problem: The task of assigning colors to vertices of

a graph so that no two adjacent vertices share the same color.
It is a classic combinatorial problem with scheduling, networking, and
optimization applications.
Determining the minimum number of colors (chromatic number) for
a graph is NP-Hard , making it computationally challenging.
Our Focus is to analyze the complexity of related problems in specific
classes of graphs and explore whether it is NP-Complete.

Akash Tayal Complexity Analysis October 5, 2024 4 / 29

Introduction

Graph Coloring Problem: The task of assigning colors to vertices of

a graph so that no two adjacent vertices share the same color.
It is a classic combinatorial problem with scheduling, networking, and
optimization applications.
Determining the minimum number of colors (chromatic number) for
a graph is NP-Hard , making it computationally challenging.
Our Focus is to analyze the complexity of related problems in specific
classes of graphs and explore whether it is NP-Complete.

Akash Tayal Complexity Analysis October 5, 2024 4 / 29

Introduction

Graph Coloring Problem: The task of assigning colors to vertices of

a graph so that no two adjacent vertices share the same color.
It is a classic combinatorial problem with scheduling, networking, and
optimization applications.
Determining the minimum number of colors (chromatic number) for
a graph is NP-Hard , making it computationally challenging.
Our Focus is to analyze the complexity of related problems in specific
classes of graphs and explore whether it is NP-Complete.

Akash Tayal Complexity Analysis October 5, 2024 4 / 29

Outline

1 Introduction

2 Preliminaries

3 Complexity of Graph Coloring Problem

4 Future Work

5 References

Akash Tayal Complexity Analysis October 5, 2024 5 / 29

Preliminaries
DEFINITION 2.1 (GRAPH)
A graph G = (V , E ) consists of V , a non-empty set of vertices(or nodes)
and E ⊆ V × V , a non-empty set of edges which have at least one vertex
associated with it, called as endpoints.

There are various types of graphs such as

D r rC D r rC
Cr
rE
@@
@@
@
@ @
r @r r @r r rB
A B A B A
Simple Complete Connected

Akash Tayal Complexity Analysis October 5, 2024 6 / 29

Preliminaries
DEFINITION 2.1 (GRAPH)
A graph G = (V , E ) consists of V , a non-empty set of vertices(or nodes)
and E ⊆ V × V , a non-empty set of edges which have at least one vertex
associated with it, called as endpoints.

There are various types of graphs such as

D r rC D r rC
Cr
rE
@@
@@
@
@ @
r @r r @r r rB
A B A B A
Simple Complete Connected

Akash Tayal Complexity Analysis October 5, 2024 6 / 29

Preliminaries
DEFINITION 2.1 (GRAPH)
A graph G = (V , E ) consists of V , a non-empty set of vertices(or nodes)
and E ⊆ V × V , a non-empty set of edges which have at least one vertex
associated with it, called as endpoints.

There are various types of graphs such as

D r rC D r rC
Cr
rE
@@
@@
@
@ @
r @r r @r r rB
A B A B A
Simple Complete Connected

Akash Tayal Complexity Analysis October 5, 2024 6 / 29

Preliminaries
DEFINITION 2.1 (GRAPH)
A graph G = (V , E ) consists of V , a non-empty set of vertices(or nodes)
and E ⊆ V × V , a non-empty set of edges which have at least one vertex
associated with it, called as endpoints.

There are various types of graphs such as

D r rC D r rC
Cr
rE
@@
@@
@
@ @
r @r r @r r rB
A B A B A
Simple Complete Connected

Akash Tayal Complexity Analysis October 5, 2024 6 / 29

Preliminaries
DEFINITION 2.1 (GRAPH)
A graph G = (V , E ) consists of V , a non-empty set of vertices(or nodes)
and E ⊆ V × V , a non-empty set of edges which have at least one vertex
associated with it, called as endpoints.

There are various types of graphs such as

D r rC D r rC
Cr
rE
@@
@@
@
@ @
r @r r @r r rB
A B A B A
Simple Complete Connected

Akash Tayal Complexity Analysis October 5, 2024 6 / 29

Preliminaries

DEFINITION 2.2 (DEGREE)

Given a graph G = (V , E ), then the degree of a vertex v ∈ V (G) is
defined as d(v ) = |N(v )|, where, N(v ) is the set of neighbors, defined as
N(v ) = {u ∈ V (G) : uv ∈ E (G)}

DEFINITIONS 2.3 (CLIQUE)

Given a graph G = (V , E ), for any C ⊆ V , if every pair of distinct vertices
in C are adjacent in G[C ](subgraph induced by G on C ) then C is known
as clique.

Akash Tayal Complexity Analysis October 5, 2024 7 / 29

Preliminaries

DEFINITION 2.2 (DEGREE)

Given a graph G = (V , E ), then the degree of a vertex v ∈ V (G) is
defined as d(v ) = |N(v )|, where, N(v ) is the set of neighbors, defined as
N(v ) = {u ∈ V (G) : uv ∈ E (G)}

DEFINITIONS 2.3 (CLIQUE)

Given a graph G = (V , E ), for any C ⊆ V , if every pair of distinct vertices
in C are adjacent in G[C ](subgraph induced by G on C ) then C is known
as clique.

Akash Tayal Complexity Analysis October 5, 2024 7 / 29

Preliminaries

DEFINITION 2.4 (GRAPH COLORING)

Given a graph G = (V , E ), a k-Coloring of a graph G is the assignment
of colors to the vertices of graph, i.e., f : V (G) → {1, 2, 3, . . . , k}.
Moreover, the vertices of one color form a color class.

There are various types of graph coloring such as

Vertex Coloring Edge Coloring Exact p-Coloring (p=2)

Akash Tayal Complexity Analysis October 5, 2024 8 / 29

Preliminaries

DEFINITION 2.4 (GRAPH COLORING)

Given a graph G = (V , E ), a k-Coloring of a graph G is the assignment
of colors to the vertices of graph, i.e., f : V (G) → {1, 2, 3, . . . , k}.
Moreover, the vertices of one color form a color class.

There are various types of graph coloring such as

Vertex Coloring Edge Coloring Exact p-Coloring (p=2)

Akash Tayal Complexity Analysis October 5, 2024 8 / 29

Preliminaries

DEFINITION 2.4 (GRAPH COLORING)

Given a graph G = (V , E ), a k-Coloring of a graph G is the assignment
of colors to the vertices of graph, i.e., f : V (G) → {1, 2, 3, . . . , k}.
Moreover, the vertices of one color form a color class.

There are various types of graph coloring such as

Vertex Coloring Edge Coloring Exact p-Coloring (p=2)

Akash Tayal Complexity Analysis October 5, 2024 8 / 29

Preliminaries

DEFINITION 2.4 (GRAPH COLORING)

Given a graph G = (V , E ), a k-Coloring of a graph G is the assignment
of colors to the vertices of graph, i.e., f : V (G) → {1, 2, 3, . . . , k}.
Moreover, the vertices of one color form a color class.

There are various types of graph coloring such as

Vertex Coloring Edge Coloring Exact p-Coloring (p=2)

Akash Tayal Complexity Analysis October 5, 2024 8 / 29

Preliminaries

DEFINITION 2.5 (EXACT SQUARE COLORING)

Given a graph G = (V , E ), an exact square coloring of G is a function
f : V (G) → {1, 2, 3, . . . , k}, such that any two vertices at a distance
exactly 2 receive distinct colors.

DEFINITION 2.6 (EXACT SQUARE CHROMATIC NUMBER)

Given a graph G = (V , E ), an exact square chromatic number is the
minimum number of colors required by an exact square coloring, denoted
as χ[#2] (G).

Akash Tayal Complexity Analysis October 5, 2024 9 / 29

Preliminaries

DEFINITION 2.5 (EXACT SQUARE COLORING)

Given a graph G = (V , E ), an exact square coloring of G is a function
f : V (G) → {1, 2, 3, . . . , k}, such that any two vertices at a distance
exactly 2 receive distinct colors.

DEFINITION 2.6 (EXACT SQUARE CHROMATIC NUMBER)

Given a graph G = (V , E ), an exact square chromatic number is the
minimum number of colors required by an exact square coloring, denoted
as χ[#2] (G).

Akash Tayal Complexity Analysis October 5, 2024 9 / 29

Outline

1 Introduction

2 Preliminaries

3 Complexity of Graph Coloring Problem

4 Future Work

5 References

Akash Tayal Complexity Analysis October 5, 2024 10 / 29

Complexity of Graph Coloring Problem
DEFINITION 3.1 (REDUCTION)
Let us consider a problem π1 (decision problem) such that the problem is
polynomially reducible to π2 if ∃ f: I(π1 )→ I(π2 ) such that;
f can be computed in polynomial time.
I ∈ Y(I(π1 )) ↔ f(I) ∈ Y(I(π2 )).
it is denoted as π1 ≤p π2 , where I is an instance.

Instance for π1 Instance for π2

Polynomial time re-
duction algorithm

Figure. Reduction Process

Akash Tayal Complexity Analysis October 5, 2024 11 / 29

Complexity of Graph Coloring Problem
DEFINITION 3.1 (REDUCTION)
Let us consider a problem π1 (decision problem) such that the problem is
polynomially reducible to π2 if ∃ f: I(π1 )→ I(π2 ) such that;
f can be computed in polynomial time.
I ∈ Y(I(π1 )) ↔ f(I) ∈ Y(I(π2 )).
it is denoted as π1 ≤p π2 , where I is an instance.

Instance for π1 Instance for π2

Polynomial time re-
duction algorithm

Figure. Reduction Process

Akash Tayal Complexity Analysis October 5, 2024 11 / 29

Complexity of Graph Coloring Problem
DEFINITION 3.2 (P and NP)
P: It consists of those problems that are solvable in polynomial time.
NP: It consists of those problems that are ‘verifiable’ in polynomial
time.
Polynomial time is the time required to solve a problem grows at a
polynomial rate relative to the size of the input.

DEFINITION 3.3 (NP-Complete)

A problem π is said to be NP-Complete if
π ∈ NP.
π 0 ≤p π , ∀ π 0 ∈ NP.

DEFINITION 3.4 (NP-Hard)

A problem π is said to be NP-Hard if π 0 ≤p π , ∀ π 0 ∈ NP
Akash Tayal Complexity Analysis October 5, 2024 12 / 29
Complexity of Graph Coloring Problem
DEFINITION 3.2 (P and NP)
P: It consists of those problems that are solvable in polynomial time.
NP: It consists of those problems that are ‘verifiable’ in polynomial
time.
Polynomial time is the time required to solve a problem grows at a
polynomial rate relative to the size of the input.

DEFINITION 3.3 (NP-Complete)

A problem π is said to be NP-Complete if
π ∈ NP.
π 0 ≤p π , ∀ π 0 ∈ NP.

DEFINITION 3.4 (NP-Hard)

A problem π is said to be NP-Hard if π 0 ≤p π , ∀ π 0 ∈ NP
Akash Tayal Complexity Analysis October 5, 2024 12 / 29
Complexity of Graph Coloring Problem
DEFINITION 3.2 (P and NP)
P: It consists of those problems that are solvable in polynomial time.
NP: It consists of those problems that are ‘verifiable’ in polynomial
time.
Polynomial time is the time required to solve a problem grows at a
polynomial rate relative to the size of the input.

DEFINITION 3.3 (NP-Complete)

A problem π is said to be NP-Complete if
π ∈ NP.
π 0 ≤p π , ∀ π 0 ∈ NP.

DEFINITION 3.4 (NP-Hard)

A problem π is said to be NP-Hard if π 0 ≤p π , ∀ π 0 ∈ NP
Akash Tayal Complexity Analysis October 5, 2024 12 / 29
Theorem 3.1
A graph G = (V , E ) is an undirected path graph if and only if there exists
a tree T (a connected a-cyclic graph) with V (T ) = C(G), where C(G) is
the set of all maximal cliques of G, such that T [Cv (G)] is a path in T for
each v ∈ V (G), Cv (G) is the set of all maximal cliques of G containing
the vertex v .

Theorem 3.2
Exact Square Coloring Problem is NP-Complete for undirected graphs.

Proof
For a vertex coloring c of an undirected path graph G, it can be easily
verified in polynomial time whether c is an exact square coloring of G.
Hence ESCP is in NP for undirected path graphs. Now we provide a
polynomial reduction from the proper coloring problem which is known to
be NP-Complete for general graphs.

Akash Tayal Complexity Analysis October 5, 2024 13 / 29

Theorem 3.1
A graph G = (V , E ) is an undirected path graph if and only if there exists
a tree T (a connected a-cyclic graph) with V (T ) = C(G), where C(G) is
the set of all maximal cliques of G, such that T [Cv (G)] is a path in T for
each v ∈ V (G), Cv (G) is the set of all maximal cliques of G containing
the vertex v .

Theorem 3.2
Exact Square Coloring Problem is NP-Complete for undirected graphs.

Proof
For a vertex coloring c of an undirected path graph G, it can be easily
verified in polynomial time whether c is an exact square coloring of G.
Hence ESCP is in NP for undirected path graphs. Now we provide a
polynomial reduction from the proper coloring problem which is known to
be NP-Complete for general graphs.

Akash Tayal Complexity Analysis October 5, 2024 13 / 29

Theorem 3.1
A graph G = (V , E ) is an undirected path graph if and only if there exists
a tree T (a connected a-cyclic graph) with V (T ) = C(G), where C(G) is
the set of all maximal cliques of G, such that T [Cv (G)] is a path in T for
each v ∈ V (G), Cv (G) is the set of all maximal cliques of G containing
the vertex v .

Theorem 3.2
Exact Square Coloring Problem is NP-Complete for undirected graphs.

Proof
For a vertex coloring c of an undirected path graph G, it can be easily
verified in polynomial time whether c is an exact square coloring of G.
Hence ESCP is in NP for undirected path graphs. Now we provide a
polynomial reduction from the proper coloring problem which is known to
be NP-Complete for general graphs.

Akash Tayal Complexity Analysis October 5, 2024 13 / 29

Theorem 3.1
A graph G = (V , E ) is an undirected path graph if and only if there exists
a tree T (a connected a-cyclic graph) with V (T ) = C(G), where C(G) is
the set of all maximal cliques of G, such that T [Cv (G)] is a path in T for
each v ∈ V (G), Cv (G) is the set of all maximal cliques of G containing
the vertex v .

Theorem 3.2
Exact Square Coloring Problem is NP-Complete for undirected graphs.

Proof
For a vertex coloring c of an undirected path graph G, it can be easily
verified in polynomial time whether c is an exact square coloring of G.
Hence ESCP is in NP for undirected path graphs. Now we provide a
polynomial reduction from the proper coloring problem which is known to
be NP-Complete for general graphs.

Akash Tayal Complexity Analysis October 5, 2024 13 / 29

Polynomial reduction
Subdivide each edge, ei ∈ E (G) to get new vertices vei .
Add edges vei vej for each i ̸= j, 1 ≤ i, j ≤ m.
Add a vertex x ∗ and add edges x ∗ vei for each ei ∈ E (G).

Figure: Construction of H from G and its clique tree

Akash Tayal Complexity Analysis October 5, 2024 14 / 29

Claim 3.2.1. The constructed graph H is an undirected path graph
We will provide a clique tree T with the vertices of T corresponding to
the maximal cliques of H and we prove that each vertex of H, the
vertices of T corresponding to the maximal clique containing that vertex
forms a path.
Cvi is the only maximal clique contains vi hence forms a trivial path.
Clearly ve appears in exactly three maximal cliques, which forms a
path in T .
At last Cx ∗ is the only maximal clique contains x ∗ .
Then by theorem 3.1 we can say H is an undirected path graph

Akash Tayal Complexity Analysis October 5, 2024 15 / 29

Claim 3.2.1. The constructed graph H is an undirected path graph
We will provide a clique tree T with the vertices of T corresponding to
the maximal cliques of H and we prove that each vertex of H, the
vertices of T corresponding to the maximal clique containing that vertex
forms a path.
Cvi is the only maximal clique contains vi hence forms a trivial path.
Clearly ve appears in exactly three maximal cliques, which forms a
path in T .
At last Cx ∗ is the only maximal clique contains x ∗ .
Then by theorem 3.1 we can say H is an undirected path graph

Akash Tayal Complexity Analysis October 5, 2024 15 / 29

Claim 3.2.1. The constructed graph H is an undirected path graph
We will provide a clique tree T with the vertices of T corresponding to
the maximal cliques of H and we prove that each vertex of H, the
vertices of T corresponding to the maximal clique containing that vertex
forms a path.
Cvi is the only maximal clique contains vi hence forms a trivial path.
Clearly ve appears in exactly three maximal cliques, which forms a
path in T .
At last Cx ∗ is the only maximal clique contains x ∗ .
Then by theorem 3.1 we can say H is an undirected path graph

Akash Tayal Complexity Analysis October 5, 2024 15 / 29

Claim 3.2.1. The constructed graph H is an undirected path graph
We will provide a clique tree T with the vertices of T corresponding to
the maximal cliques of H and we prove that each vertex of H, the
vertices of T corresponding to the maximal clique containing that vertex
forms a path.
Cvi is the only maximal clique contains vi hence forms a trivial path.
Clearly ve appears in exactly three maximal cliques, which forms a
path in T .
At last Cx ∗ is the only maximal clique contains x ∗ .
Then by theorem 3.1 we can say H is an undirected path graph

Akash Tayal Complexity Analysis October 5, 2024 15 / 29

Claim 3.2.1. The constructed graph H is an undirected path graph
We will provide a clique tree T with the vertices of T corresponding to
the maximal cliques of H and we prove that each vertex of H, the
vertices of T corresponding to the maximal clique containing that vertex
forms a path.
Cvi is the only maximal clique contains vi hence forms a trivial path.
Clearly ve appears in exactly three maximal cliques, which forms a
path in T .
At last Cx ∗ is the only maximal clique contains x ∗ .
Then by theorem 3.1 we can say H is an undirected path graph

Akash Tayal Complexity Analysis October 5, 2024 15 / 29

Claim 3.2.2. χ(G) ≤ k if and only if χ[#2] (G) ≤ k + 1

Proof
Given a function f: V(G) → [k], define ϕ : V (H) → [k + 1] as follows:
ϕ(vi ) = f (vi ) for each 1 ≤ i ≤ n.
ϕ(vei ) = ϕ(x ∗ ) = c ∗ for each 1 ≤ i ≤ m.
Any pair of vertices that are adjacent in G are at a distance 2 in H
and receive distinct colors.
Also x ∗ is adjacent to all vi so they receive distinct colors.
Also vei , vej and x ∗ are adjacent in H so they receive same color as c ∗ .
Hence ϕ is an exact square coloring of H with k + 1 colors.

Akash Tayal Complexity Analysis October 5, 2024 16 / 29

Claim 3.2.2. χ(G) ≤ k if and only if χ[#2] (G) ≤ k + 1

Proof
Given a function f: V(G) → [k], define ϕ : V (H) → [k + 1] as follows:
ϕ(vi ) = f (vi ) for each 1 ≤ i ≤ n.
ϕ(vei ) = ϕ(x ∗ ) = c ∗ for each 1 ≤ i ≤ m.
Any pair of vertices that are adjacent in G are at a distance 2 in H
and receive distinct colors.
Also x ∗ is adjacent to all vi so they receive distinct colors.
Also vei , vej and x ∗ are adjacent in H so they receive same color as c ∗ .
Hence ϕ is an exact square coloring of H with k + 1 colors.

Akash Tayal Complexity Analysis October 5, 2024 16 / 29

Claim 3.2.2. χ(G) ≤ k if and only if χ[#2] (G) ≤ k + 1

Proof
Given a function f: V(G) → [k], define ϕ : V (H) → [k + 1] as follows:
ϕ(vi ) = f (vi ) for each 1 ≤ i ≤ n.
ϕ(vei ) = ϕ(x ∗ ) = c ∗ for each 1 ≤ i ≤ m.
Any pair of vertices that are adjacent in G are at a distance 2 in H
and receive distinct colors.
Also x ∗ is adjacent to all vi so they receive distinct colors.
Also vei , vej and x ∗ are adjacent in H so they receive same color as c ∗ .
Hence ϕ is an exact square coloring of H with k + 1 colors.

Akash Tayal Complexity Analysis October 5, 2024 16 / 29

Claim 3.2.2. χ(G) ≤ k if and only if χ[#2] (G) ≤ k + 1

Proof
Given a function f: V(G) → [k], define ϕ : V (H) → [k + 1] as follows:
ϕ(vi ) = f (vi ) for each 1 ≤ i ≤ n.
ϕ(vei ) = ϕ(x ∗ ) = c ∗ for each 1 ≤ i ≤ m.
Any pair of vertices that are adjacent in G are at a distance 2 in H
and receive distinct colors.
Also x ∗ is adjacent to all vi so they receive distinct colors.
Also vei , vej and x ∗ are adjacent in H so they receive same color as c ∗ .
Hence ϕ is an exact square coloring of H with k + 1 colors.

Akash Tayal Complexity Analysis October 5, 2024 16 / 29

Claim 3.2.2. χ(G) ≤ k if and only if χ[#2] (G) ≤ k + 1

Proof
Given a function f: V(G) → [k], define ϕ : V (H) → [k + 1] as follows:
ϕ(vi ) = f (vi ) for each 1 ≤ i ≤ n.
ϕ(vei ) = ϕ(x ∗ ) = c ∗ for each 1 ≤ i ≤ m.
Any pair of vertices that are adjacent in G are at a distance 2 in H
and receive distinct colors.
Also x ∗ is adjacent to all vi so they receive distinct colors.
Also vei , vej and x ∗ are adjacent in H so they receive same color as c ∗ .
Hence ϕ is an exact square coloring of H with k + 1 colors.

Akash Tayal Complexity Analysis October 5, 2024 16 / 29

Proof of the converse part
Given an exact square coloring ϕ of H using k + 1.
Any vertex of V can not receive the same color as they are at
distance 2 in H and ϕ is exact square coloring.
So vertices of V are assigned at most k colors.
Hence the result holds

Akash Tayal Complexity Analysis October 5, 2024 17 / 29

Proof of the converse part
Given an exact square coloring ϕ of H using k + 1.
Any vertex of V can not receive the same color as they are at
distance 2 in H and ϕ is exact square coloring.
So vertices of V are assigned at most k colors.
Hence the result holds

Akash Tayal Complexity Analysis October 5, 2024 17 / 29

Proof of the converse part
Given an exact square coloring ϕ of H using k + 1.
Any vertex of V can not receive the same color as they are at
distance 2 in H and ϕ is exact square coloring.
So vertices of V are assigned at most k colors.
Hence the result holds

Akash Tayal Complexity Analysis October 5, 2024 17 / 29

Proof of the converse part
Given an exact square coloring ϕ of H using k + 1.
Any vertex of V can not receive the same color as they are at
distance 2 in H and ϕ is exact square coloring.
So vertices of V are assigned at most k colors.
Hence the result holds

Akash Tayal Complexity Analysis October 5, 2024 17 / 29

Definitions
Simplicial Vertex: A vertex v of a given graph G = (V , E ) is called
simplicial if NG (v ) is a clique.
Perfect Elimination Ordering(PEO): Let α = (v1 , v2 , . . . , vn ) be an
ordering of G. We say that α is a PEO of G if vi is a simplicial vertex
of Gi = G[{vi , vi+1 , . . . , vn }] for all 1 ≤ i ≤ n.
Bi-Compatible elimination Ordering(BCO): A perfect elimination
ordering α = (v1 , v2 , . . . , vn ) is bi-compatible elimination ordering of
G if α−1 = (vn , vn−1 , . . . , v1 ) is also a PEO of G.

Theorem 3.3
A graph G has a bi-compatible elimination ordering if and only if G is a
proper interval graph.

Akash Tayal Complexity Analysis October 5, 2024 18 / 29

Definitions
Simplicial Vertex: A vertex v of a given graph G = (V , E ) is called
simplicial if NG (v ) is a clique.
Perfect Elimination Ordering(PEO): Let α = (v1 , v2 , . . . , vn ) be an
ordering of G. We say that α is a PEO of G if vi is a simplicial vertex
of Gi = G[{vi , vi+1 , . . . , vn }] for all 1 ≤ i ≤ n.
Bi-Compatible elimination Ordering(BCO): A perfect elimination
ordering α = (v1 , v2 , . . . , vn ) is bi-compatible elimination ordering of
G if α−1 = (vn , vn−1 , . . . , v1 ) is also a PEO of G.

Theorem 3.3
A graph G has a bi-compatible elimination ordering if and only if G is a
proper interval graph.

Akash Tayal Complexity Analysis October 5, 2024 18 / 29

Definitions
Simplicial Vertex: A vertex v of a given graph G = (V , E ) is called
simplicial if NG (v ) is a clique.
Perfect Elimination Ordering(PEO): Let α = (v1 , v2 , . . . , vn ) be an
ordering of G. We say that α is a PEO of G if vi is a simplicial vertex
of Gi = G[{vi , vi+1 , . . . , vn }] for all 1 ≤ i ≤ n.
Bi-Compatible elimination Ordering(BCO): A perfect elimination
ordering α = (v1 , v2 , . . . , vn ) is bi-compatible elimination ordering of
G if α−1 = (vn , vn−1 , . . . , v1 ) is also a PEO of G.

Theorem 3.3
A graph G has a bi-compatible elimination ordering if and only if G is a
proper interval graph.

Akash Tayal Complexity Analysis October 5, 2024 18 / 29

Definitions
Simplicial Vertex: A vertex v of a given graph G = (V , E ) is called
simplicial if NG (v ) is a clique.
Perfect Elimination Ordering(PEO): Let α = (v1 , v2 , . . . , vn ) be an
ordering of G. We say that α is a PEO of G if vi is a simplicial vertex
of Gi = G[{vi , vi+1 , . . . , vn }] for all 1 ≤ i ≤ n.
Bi-Compatible elimination Ordering(BCO): A perfect elimination
ordering α = (v1 , v2 , . . . , vn ) is bi-compatible elimination ordering of
G if α−1 = (vn , vn−1 , . . . , v1 ) is also a PEO of G.

Theorem 3.3
A graph G has a bi-compatible elimination ordering if and only if G is a
proper interval graph.

Akash Tayal Complexity Analysis October 5, 2024 18 / 29

Lemma 3.1
Let G be a proper interval graph with BCO, α = (v1 , v2 , . . . , vn ). Then
for any vi ∈ G,
if vj , vk ∈ NG2 (vi ) such that i ≤ j ≤ k, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.
if vj , vk ∈ NG2 (vi ) such that j ≤ k ≤ i, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.

Theorem 3.4
Algorithm 1 returns χ[#2] (G) and an optimal exact square coloring of a
proper interval graph G in O(n + m).

Proof: Let G = (V , E ) be any connected proper interval graph such that

G ̸= Kn . Then, χ[#2] (G) ≥ 2. Let m(i) and l(i) denote the index of the
highest and least index neighbor respectively.

Akash Tayal Complexity Analysis October 5, 2024 19 / 29

Lemma 3.1
Let G be a proper interval graph with BCO, α = (v1 , v2 , . . . , vn ). Then
for any vi ∈ G,
if vj , vk ∈ NG2 (vi ) such that i ≤ j ≤ k, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.
if vj , vk ∈ NG2 (vi ) such that j ≤ k ≤ i, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.

Theorem 3.4
Algorithm 1 returns χ[#2] (G) and an optimal exact square coloring of a
proper interval graph G in O(n + m).

Proof: Let G = (V , E ) be any connected proper interval graph such that

G ̸= Kn . Then, χ[#2] (G) ≥ 2. Let m(i) and l(i) denote the index of the
highest and least index neighbor respectively.

Akash Tayal Complexity Analysis October 5, 2024 19 / 29

Lemma 3.1
Let G be a proper interval graph with BCO, α = (v1 , v2 , . . . , vn ). Then
for any vi ∈ G,
if vj , vk ∈ NG2 (vi ) such that i ≤ j ≤ k, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.
if vj , vk ∈ NG2 (vi ) such that j ≤ k ≤ i, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.

Theorem 3.4
Algorithm 1 returns χ[#2] (G) and an optimal exact square coloring of a
proper interval graph G in O(n + m).

Proof: Let G = (V , E ) be any connected proper interval graph such that

G ̸= Kn . Then, χ[#2] (G) ≥ 2. Let m(i) and l(i) denote the index of the
highest and least index neighbor respectively.

Akash Tayal Complexity Analysis October 5, 2024 19 / 29

Lemma 3.1
Let G be a proper interval graph with BCO, α = (v1 , v2 , . . . , vn ). Then
for any vi ∈ G,
if vj , vk ∈ NG2 (vi ) such that i ≤ j ≤ k, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.
if vj , vk ∈ NG2 (vi ) such that j ≤ k ≤ i, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.

Theorem 3.4
Algorithm 1 returns χ[#2] (G) and an optimal exact square coloring of a
proper interval graph G in O(n + m).

Proof: Let G = (V , E ) be any connected proper interval graph such that

G ̸= Kn . Then, χ[#2] (G) ≥ 2. Let m(i) and l(i) denote the index of the
highest and least index neighbor respectively.

Akash Tayal Complexity Analysis October 5, 2024 19 / 29

Lemma 3.1
Let G be a proper interval graph with BCO, α = (v1 , v2 , . . . , vn ). Then
for any vi ∈ G,
if vj , vk ∈ NG2 (vi ) such that i ≤ j ≤ k, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.
if vj , vk ∈ NG2 (vi ) such that j ≤ k ≤ i, then vr ∈ NG2 (vi ) for every r
with j ≤ r ≤ k.

Theorem 3.4
Algorithm 1 returns χ[#2] (G) and an optimal exact square coloring of a
proper interval graph G in O(n + m).

Proof: Let G = (V , E ) be any connected proper interval graph such that

G ̸= Kn . Then, χ[#2] (G) ≥ 2. Let m(i) and l(i) denote the index of the
highest and least index neighbor respectively.

Akash Tayal Complexity Analysis October 5, 2024 19 / 29

Claim 3.4.1 H = G [#2]
Let α = (v1 , v2 , . . . , vn ) be an BCO of G.
Clearly the vertex set is same for both.

Akash Tayal Complexity Analysis October 5, 2024 20 / 29

Continued
For any vi ∈ V (G), A(i) and B(i) correspond to m(i) and m(m(i))
This lemma shows that the vertices at a distance 2 from any given
vertex in a proper interval graph are consecutive in the BCO ordering.
The construction of G [#2] is done by maintaining arrays A(i) and
B(i), ensuring that the graph is built efficiently.

Case 1. When H is bipartite.

Figure. An example of proper interval graph G and H, obtained from Algorithm 1, where H is a bipartite graph.

Akash Tayal Complexity Analysis October 5, 2024 21 / 29

Case 2. When H is not bipartite.

Figure. An example of proper interval graph G where H is not bipartite. The ESC obtained by algorithm 1 is on right.

Claim 3.4.2 Algorithms 1 return an ESC of G using c = χ[#2] (G).

If H is bipartite, the algorithms outputs c = 2 and f as the proper
2-coloring of H, since χ[#2] (G) = χ(H) = 2.
If H is not bipartite,
Considered any two vertices vi and vk such that dG (vi , vk ) = 2.
The algorithm ensures that these vertices do not receive the same
color. If they do, it leads to a contradiction regarding their positions
and the color assignments made by the algorithm.

Akash Tayal Complexity Analysis October 5, 2024 22 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Claim 3.4.3 f can be computed in O(n + m)time.
For any vertex vi , the entry A(i) can be computed in O(d(vi )) time.
Array A can be obtained in at most O(n + m)time.
Array B obtained by accessing array A twice.
To construct H = G [#2] , we add at most O(d(vm(i) ) + d(vl(i) )) edges
for each vertex.
We can verify H is bipartite in linear time.
Further a proper 2-coloring of any bipartite graph can be obtained in
linear time.
When H is not bipartite, each vertex is processed exactly once.
Hence the results holds.

Akash Tayal Complexity Analysis October 5, 2024 23 / 29

Outline

1 Introduction

2 Preliminaries

3 Complexity of Graph Coloring Problem

4 Future Work

5 References

Akash Tayal Complexity Analysis October 5, 2024 24 / 29

Future Work

Then we have that Exact Square Coloring Problem is NP-complete

for undirected path graphs.
Also we have that there is a linear time algorithm to find the optimal
exact square coloring of proper interval graphs, O(m + n).
Investigate the identification of NP-complete problems within the
domain of graph coloring.
Explore NP-completeness results for the Exact Square Coloring
problem, particularly for subclasses of chordal graphs.
Develop and analyze algorithmic solutions for the Exact Square
Coloring problem.

Akash Tayal Complexity Analysis October 5, 2024 25 / 29

Future Work

Then we have that Exact Square Coloring Problem is NP-complete

for undirected path graphs.
Also we have that there is a linear time algorithm to find the optimal
exact square coloring of proper interval graphs, O(m + n).
Investigate the identification of NP-complete problems within the
domain of graph coloring.
Explore NP-completeness results for the Exact Square Coloring
problem, particularly for subclasses of chordal graphs.
Develop and analyze algorithmic solutions for the Exact Square
Coloring problem.

Akash Tayal Complexity Analysis October 5, 2024 25 / 29

Future Work

Then we have that Exact Square Coloring Problem is NP-complete

for undirected path graphs.
Also we have that there is a linear time algorithm to find the optimal
exact square coloring of proper interval graphs, O(m + n).
Investigate the identification of NP-complete problems within the
domain of graph coloring.
Explore NP-completeness results for the Exact Square Coloring
problem, particularly for subclasses of chordal graphs.
Develop and analyze algorithmic solutions for the Exact Square
Coloring problem.

Akash Tayal Complexity Analysis October 5, 2024 25 / 29

Future Work

Then we have that Exact Square Coloring Problem is NP-complete

for undirected path graphs.
Also we have that there is a linear time algorithm to find the optimal
exact square coloring of proper interval graphs, O(m + n).
Investigate the identification of NP-complete problems within the
domain of graph coloring.
Explore NP-completeness results for the Exact Square Coloring
problem, particularly for subclasses of chordal graphs.
Develop and analyze algorithmic solutions for the Exact Square
Coloring problem.

Akash Tayal Complexity Analysis October 5, 2024 25 / 29

Future Work

Then we have that Exact Square Coloring Problem is NP-complete

for undirected path graphs.
Also we have that there is a linear time algorithm to find the optimal
exact square coloring of proper interval graphs, O(m + n).
Investigate the identification of NP-complete problems within the
domain of graph coloring.
Explore NP-completeness results for the Exact Square Coloring
problem, particularly for subclasses of chordal graphs.
Develop and analyze algorithmic solutions for the Exact Square
Coloring problem.

Akash Tayal Complexity Analysis October 5, 2024 25 / 29

Outline

1 Introduction

2 Preliminaries

3 Complexity of Graph Coloring Problem

4 Future Work

5 References

Akash Tayal Complexity Analysis October 5, 2024 26 / 29

Reference

Thomas H Cormen, Charles E Leiserson, Ronald L Rivest, Clifford

Stein:Introduction to Algorithms, MIT Press, (2001)

Douglas Brent West : Introduction to Graph Theory Secomd

Edition,Prentice Hall, New York,(2001)

Akash Tayal Complexity Analysis October 5, 2024 27 / 29

Reference

[1] Priyamvada and B. S. Panda: Exact square coloring of certain classes

of graphs: Complexity and algorithms. Theoretical Computer Science,
932, 84-101, 2022.
[2] Florent Foucaud, Hervé Hocquard, Suchismita Mishra, Narayanan
Narayanan, Reza Naserasr, Eric Sopena, Petru Valicov Exact square
coloring of subcubic planar graphs Discrete Appl. Math., 293 (2021),
74-89
[3] M.A. Shalu, S. Vijayakumar, S. Devi Yamini, T.P. Sandhya, On the
algorithmic aspects of strong subcoloring, J. Comb. Optim., 35 (4)
(2018), 1312-1329

Akash Tayal Complexity Analysis October 5, 2024 28 / 29

 Thank You !

Akash Tayal Complexity Analysis October 5, 2024 29 / 29