Exploration of Kruskal's Algorithm in Graph Theory

Abstract

This paper presents a comprehensive exploration of Kruskal's algorithm, a fundamental method in graph theory and
optimization, aimed at finding the minimum spanning tree (MST) in connected weighted graphs. It delves into the algorithm's core
principles and step-by-step execution, shedding light on its key attributes and relevance in a variety of real-world applications. The
primary motivation for this study lies in the algorithm's efficiency in producing minimal spanning trees and its adaptability for
addressing various practical problems, ranging from network design to circuit routing. Through a detailed analysis of Kruskal's
algorithm, this paper provides insights into its time complexity and correctness. Additionally, it provides empirical results,
demonstrating the algorithm's effectiveness in producing optimal MST solutions, making it an indispensable tool for optimization
and connectivity challenges in diverse domains.

Keywords: Graph Theory; Kruskal's Algorithm; Minimum Spanning Tree

1. Introduction
1.1. Greedy Algorithm
A greedy algorithm is a powerful and efficient problem-solving approach known for its simplicity and
effectiveness. It excels in making the best immediate choice consistently leads to an optimal global solution. This
approach simplifies complex problems, making them more manageable and computationally efficient, often
providing elegant solutions to a wide range of real-world challenges. The complexity of a greedy algorithm is
typically linear or close to linear, O(n) or O(n log n), making it efficient for many problems.

1.2. Minimum spanning tree

A minimum spanning tree (MST) is effectively discovered through the application of a greedy approach, in
algorithms like Kruskal's and Prim's. These algorithms aim to connect all nodes with the least cumulative edge
weight while avoiding cycles. They consistently make locally optimal choices by favoring edges with the smallest
weights at each step. MSTs have diverse practical applications across fields such as network design and clustering.
They play a crucial role in efficiently allocating resources and planning cost-effective infrastructure. The core
principle of these algorithms is to progressively build a tree by prioritizing the lightest edges, ensuring a connected,
acyclic structure with minimal total weight, demonstrating the versatility and efficiency of the greedy approach.
In a connected undirected graph, denoted as U=(M,N), the formal definition of a spanning tree is represented as U
′=(M,N′), where N′⊆N. For a connected, undirected weighted graph U=(M,N,w), the minimum spanning tree (MST)
is characterized by having the minimum weight. The weight or cost (c(t)) of the tree T is precisely defined as the
summation of the weights of its edges which was expressed in [15] as,

c(t)=∑n∈(T)c(n) (1)

1.3. Kruskal’s algorithm

Kruskal's algorithm is a widely applied technique in graph theory for determining the Minimum Spanning Tree
(MST) in a connected, weighted graph. The algorithm initiates by treating each vertex as an isolated component. It
then proceeds to examine the edges of the graph, sorting them in ascending order of their weights. At each step, it
assesses whether incorporating the current edge into the MST would lead to a cycle. If this is not the case, the edge is
included in the MST, and the two components it connects are merged into one. This process continues until the MST
is complete, with all vertices connected and no cycles present. Kruskal's algorithm is both efficient and capable of
delivering a globally optimal solution, particularly for sparse graphs as discussed in [3]. By emphasizing the
importance of edge weights and ensuring the resultant tree remains acyclic, Kruskal's algorithm exemplifies the
effectiveness of the greedy approach in addressing intricate graph-related problems. In Fig. 1 example problem is
shown for implementing the algorithm.
 Fig. 1. Problem

Step 1: Arrange the edges in ascending order based on their weights as shown in Tab. 1.
Table 1

Edges and its weight

Edge AC AD DE AB CD BC EF DF BE BD AE
Weight 11 22 22 31 31 45 45 50 83 87 88

Step 2: Iterate through the sorted edges, starting with the lowest weight such that no cycles get formed.

Add edge AC (weight 11) to the MST, resulting in: (AC) as shown in Fig. 2.

Fig. 2.

Step 3: Add edge AD (weight 22) to the MST, resulting in: (AC), (AD) as shown in Fig. 3.

Fig. 3.

Step 4: Add edge DE (weight 22) to the MST, resulting in: (AC), (AD), (DE) as shown Fig. 4.

Fig. 4.

Step 5: Add edge AB (weight 31) to the MST, resulting in: (AC), (AD), (DE), (AB) as shown in Fig. 5.

Fig. 5.

Step 6: Since CD, BC forms a cycle, ignore it. Add edge EF (weight 45) to the MST, resulting in: (AC), (AD), (DE),
(AB), (EF) as shown in Fig. 6.
 Fig. 6.

Step 7: Since edges DF, BE, BD, AE form a cycle, ignore it.

Step 8: Since all vertices have been incorporated into the Minimum Spanning Tree (MST), we can conclude the
process.

Weight of MST = sum of all edge weights

=11+22+22+31+45

=131

2. Literature Surevey

In paper [1], they utilized Kruskal's Algorithm, Parallel algorithms, and Helper Threads to address the absence of
explicit parallelism, enhancing algorithm performance. It achieved speedups of up to 5.5times with 8 threads in
comparison to the sequential version. It enables swift traffic rerouting for network reliability.

In paper [2], the authors applied the Extended Kruskal algorithm, denoted as EKRUS, to address delay-bound
constraints. EKRUS effectively lowered time complexity when compared with the Destination Driven Multicast
Routing Algorithm (DDMC). This approach proved valuable for multicast routing in networks with arbitrary
topologies.

In paper [3], the authors examined the Kruskal and Boruvka algorithms, highlighting Kruskal's superior time
efficiency over Boruvka. This research suggests applicability in determining optimal routes for water pipe installation
in water distribution systems.

In paper [4], they compared the Kruskal and Prim algorithms for complexity analysis. While Kruskal excelled in
space complexity, Prim demonstrated superiority in time complexity. The study also proposed that Kruskal's
algorithm and graph theory could find indirect applications in financial analysis and portfolio management.

In paper [5], they applied the Kruskal algorithm to address the challenge of minimizing costs in setting up
communication network systems. It successfully simulated and computed the minimum spanning tree and the
associated expenses, and they find it useful for solving problems like the traveling salesman, postman, and network
maximum flow.

In paper [6], a combination of depth-first search and Kruskal's algorithm was employed to tackle the Multi-Period
Degree Constrained Minimum Spanning Tree (MPDCMST) problem. This approach yielded improved solutions for
both DCMST and MPDCMST, used for potential applications in power distribution networks and sensor deployment
for environmental monitoring.

In paper [7], they used Kruskal's algorithm to determine an ideal switching sequence, considering the network's
reliability under fluctuating loads. It effectively minimized power flow in distribution lines, making an optimal
strategy for power restoration in areas experiencing outages while considering load shedding requirements.

In paper [8], they compared Kruskal's algorithm and PostgreSQL's default optimization techniques for query
optimization. Comparing Kruskal's algorithm to PostgreSQL's Dynamic Programming and GEOQ algorithms,
Kruskal is feasible for optimizing extensive queries, particularly those with a substantial number of relations in
DBMS applications.

In paper [9], the K-means algorithm was enhanced with Kruskal's method for more consistent initial cluster center
selection, mitigating the variability due to different initial values. The improved K-means algorithm was then
compared to the traditional approach, demonstrating increased stability and accuracy in clustering outcomes. This
research is specifically used in gene expression data analysis.
In paper [10], Kruskal's algorithm was used to enhance time efficiency in comparison to the basic HCPP
(Hierarchical Chinese Postman Problem) and CPP algorithms. The Kruskal-based HCPP solution demonstrated a
notable average improvement, showing its applicability in tasks like snow plowing, winter gritting, and street
sweeping.

In paper [11], Kruskal's algorithm played a major role in optimizing electrical network design. By integrating
Kruskal with Vogel's approximation and a customized distribution method, the study achieved cable length reduction
and subsequently lowered network maintenance expenses. This approach was applied in the design of electrical
networks within the Democratic Republic of Congo (DRC).

In paper [12], Kruskal's algorithm was used to determine the most efficient reconfiguration across multiple time
intervals, with a primary goal of minimizing overall power loss. The algorithm proved highly effective in terms of
both optimization quality and execution speed and was applied to dynamic distribution network reconfiguration.

In paper [13], they introduce Kruskal's algorithm for optimizing information passing within networks. By integrating
it with the Multicriteria Decision-Making (MCDM) method, the research achieved rapid and highly effective results,
making it a valuable recommendation for application in social media networks.

In paper [14], they introduce Kruskal's algorithm for minimizing redundant histogram computations. When compared
to the native approach, using Kruskal's algorithm followed by depth-first traversal demonstrated superior
performance in terms of computation time and memory efficiency.

In paper [15], they used Kruskal's and Prim's algorithms to evaluate their performance on Nigeria's 36-state road
network. Kruskal's algorithm consistently generated a minimal spanning tree with shorter kilometer distance
compared to Prim's. This paper suggests Kruskal’s to be used in network design such as roads, telephone, electrical
and cable-laying.

In paper [16], they used Kruskal's algorithm to enhance clustering efficiency, comparing it with the FAST-clustering
algorithm using the probability-based Naive Bayes Classifier. Kruskal's algorithm enhanced classifier performance
and significantly reduced processing time.

In paper [17], they used Kruskal's, Prim's, and Boruvka's algorithms to assess their compatibility with various parallel
MST algorithms. Kruskal's algorithm was found to be the least favorable for parallelization. The paper suggests their
applicable in diverse fields, including wireless communication, cancer detection, and medical imaging.

In paper [18], they used Kruskal's algorithm for efficient network design, proving its simplicity and utility in
optimizing electrical networks for applications such as Rural Electrification and the cost-effective transmission line
networks.

In paper [19], Kruskal's algorithm is introduced to enhance the functional lifespan of sensor networks which reduced
the maximum energy consumption per sensor node. The research recommends its application in the context of sensor
networks.

In paper [20], the introduction of the Filter-Kruskal algorithm was proposed as a dependable method for cluster
structure retrieval. Compared to single linkage clustering MST, the Filter-Kruskal algorithm demonstrated notable
speed and precision, particularly for the customer segmentation datasets used in the research.

In paper [21], they applied Kruskal's algorithm to assess the topology of radial networks, combining it with the
Euclidean distance metric. The resulting approach exhibited strong overall performance.

In paper [22], it uses Kruskal's algorithm in hierarchical clustering to achieve Fuzzy Image Segmentation. By
combining Kruskal's algorithm with the Divide and Link method, it effectively sorts edges during the partitioning
process. The application of this approach extends to social network analysis.

In paper [23], Kruskal's algorithm was used to handle complexity in reconfiguring large-scale distribution networks.
Comparing it to the branch and bound method, Kruskal's algorithm achieved minimal losses, improved voltage
magnitudes, and less computational time. This approach was applied to the distribution network reconfiguration
problem.
In paper [24], it introduces Kruskal's algorithm for feature selection based on feature correlation calculation. A
comparison with the Markov clustering algorithm proved Kruskal's ability to achieve high classification accuracy.
This method was applied to select salient features from high-dimensional data.

Paper [25], presents Kruskal's algorithm for parallel reconfiguration problem solution, combined with genetic
algorithms for optimal capacitor allocation. The integrated approach effectively minimizes power loss and enhances
voltage profiles within the distribution network.

In paper [26], they applied Kruskal's algorithm to construct all-to-all broadcast trees, enhancing it with the r-shrink
algorithm and BIP. The combined approach effectively minimized total power consumption in broadcasting trees and
found application in wireless networks.

In paper [27], they proposed an automated method for cluster center determination and bidirectional clustering in
electricity user contexts. It utilized the Kruskal-based K-Means clustering algorithm, resulting in improved
recommendation accuracy, efficiency, and scalability. The approach was implemented in a recommendation system.

Paper [28], applied Kruskal's algorithm in the development and investigation of an agricultural economic output rate
target detection system. In comparison with Prim's algorithm, Kruskal demonstrated an average reduction in
instability rate. They implemented in agriculture target detection.

In paper [29], Kruskal's algorithm was used to achieve power balance in energy consumption, extending the battery
life of devices and mitigating energy hot spot routing challenges. This approach effectively reduced power usage in
wireless networks.

In paper [30], they used Kruskal's algorithm to optimize power supply restoration for critical loads, minimizing
switch action times and adhering to node voltage constraints. Comparing it with the genetic algorithm, Kruskal's
approach proved more efficient, finding application in distribution network reconfiguration strategies.

Tab. 2 discuss the inference of papers which we discussed above,

Table 2

Inference of surveys

Papers Inference
[1] addresses absence of explicit parallelism.
traffic rerouting for network.
[2] delay-bound constraints.
lowered time complexity.
multicast routing in networks.
[3] Kruskal's time efficiency over Boruvka.
optimal routes for water pipe installation.
[4] complexity analysis.
financial analysis and portfolio management.
[5] minimizing costs in communication network.
traveling salesman.
[6] Multi-Period Degree Constrained Minimum Spanning Tree
power distribution networks.
[7] ideal switching sequence.
optimal strategy for power restoration.
[8] query optimization.
[9] initial cluster center selection.
gene expression data analysis.
[10] enhance time efficiency.
HCPP (Hierarchical Chinese Postman Problem) and CPP algorithms.
[11] optimizing electrical network design.
cable length reduction.
lowered network maintenance expenses.
[12] minimizing power loss.
network reconfiguration.
[13] optimizing information passing in networks.
highly effective.
[14] minimizing redundant histogram computations.
memory efficiency.
[15] road network.
[16] enhance clustering efficiency.
reduced processing time.
[17] compatibility with various parallel algorithms.
[18] efficient network design.
 rural electrification.
[19] lifespan of sensor networks.
Reduced energy consumption.
[20] cluster structure retrieval.
notable speed.
customer segmentation.
[21] topology of radial networks.
[22] Fuzzy Image Segmentation.
social network analysis.
[23] large-scale distribution networks.
improved voltage magnitudes.
[24] feature selection.
high classification accuracy.
[25] parallel reconfiguration problem.
[26] all-to-all broadcast trees.
wireless networks.
[27] bidirectional clustering.
recommendation system.
[28] target detection system.
[29] power balance.
extending the battery life.
[30] optimize power supply restoration.
more efficient.

3. Algorithm

Kruskal's algorithm is a fundamental graph theory technique for finding a minimum spanning tree (MST) in a
weighted graph. It operates by iteratively selecting edges with the lowest weights while ensuring that no cycles are
formed, ultimately growing the MST until all vertices are included. To achieve this, the algorithm employs a disjoint-
set data structure, which efficiently tracks the connected components of the graph. The algorithm terminates once the
MST contains (V-1) edges, where V represents the number of vertices in the graph. Kruskal's algorithm's time
complexity is primarily dominated by the sorting of edges based on their weights, typically achieved through
quicksort or merge sort, resulting in O (E log E) time complexity, where E represents the number of edges in the
graph. Its systematic consideration of edges by weight and its ability to handle disconnected components render it a
potent tool for solving a wide spectrum of optimization challenges in fields such as network design and urban
planning. As [15] describes algorithm and notations are explained in [31] which are described below.
Algorithm 1 Algorithm
1: KRUSKAL(graph)
2: T=Empty;
3: For every node n ∈ G.N:
4: CreateSet(n)
5: For every path (m,n) ∈ G,E arranged by increasing weights(m,n):
6: If NewSet(m) ≠NewSet(n):
7: T = T ∪{(m,n)} UNION (m.n)
8: return T
Where notations include with procedure,

G (V, E) where,

V represents vertices,

E represents Edges,

E' represents set of edges sorted by non-decreasing weight.

1.Initialize an empty set T to represent the minimum spanning tree.

2.For each edge e = (u, v) in E', if adding e to T does not create a cycle in the current subgraph (formed by the edges
in T and e), add e to T.

3.Repeat this process until T contains V-1 edges, which forms the minimum spanning tree of G.

4. Applications

4.1. Early applications

Network Design: Kruskal's algorithm is a crucial tool in network design, such as telecommunications and computer
networks. It assists in selecting the most cost-efficient links or connections to ensure network connectivity while
minimizing expenses. As [11] explained that by identifying the minimum spanning tree, which includes the essential
links, network engineers can design reliable and cost-effective communication networks, ensuring that data can flow
efficiently with minimal latency.

Circuit Design: Within the field of electrical engineering and integrated circuit design, Kruskal's algorithm finds
applications in optimizing the layout of wires and connections on circuit boards. This algorithm, by identifying the
minimum spanning tree, enables the reduction of the number of wires required to connect various components, thus
simplifying the complexity of circuit layouts. This simplification, in turn, enhances the efficiency of electronic
devices.

Urban Planning: In urban planning, Kruskal's algorithm plays a pivotal role in optimizing the development of road
networks, utility connections, and infrastructure within cities and regions. By identifying the minimum spanning tree
of roads and utility lines, urban planners can effectively minimize construction costs while ensuring comprehensive
connectivity. This approach contributes to the creation of sustainable and well-organized urban environments.

4.2. Research applications

Data Centers and Cloud Computing: Within the dynamic environment of data centers and cloud computing,
Kruskal's algorithm is actively deployed in real-time scenarios to optimize server interconnections and data paths.
Through its continuous adaptation to fluctuating workloads and network conditions, the algorithm ensures efficient
data transmission, minimizing latency, and enhancing overall system performance. This real-time adaptability is
instrumental for data centers and cloud providers in meeting the diverse demands of users and applications while
conserving valuable resources.

Traffic Management and Routing: Real-time traffic management systems make use of Kruskal's algorithm to
dynamically adjust traffic routes and signal timing. By continually monitoring traffic conditions and adapting routes
to reduce congestion and travel times, it plays a crucial role in responsive and adaptive traffic control. This real-time
application is essential for alleviating traffic congestion, enhancing transportation efficiency, and enhancing the
quality of life in urban areas.

Sensor Networks: In the domain of wireless sensor networks, Kruskal's algorithm is actively employed in real-time
to establish energy-efficient spanning trees. These optimized trees streamline data collection and communication
among sensor nodes by adaptively selecting the most efficient routes. This real-time application ensures that sensor
networks operate with minimal energy consumption, extending the network's operational lifespan and enabling
continuous data collection in remote or challenging environments.

5. Experimental results

5.1. Implementation
class Graph:
def __init__(self, vertices):
self.V = vertices
self.graph = []
def addEdge(self, u, v, w):
self.graph.append([u, v, w])
def find(self, parent, i):
if parent[i] != i:
parent[i] = self.find(parent, parent[i])
return parent[i]
def union(self, parent, rank, x, y):
if rank[x] < rank[y]:
parent[x] = y
elif rank[x] > rank[y]:
parent[y] = x
else:
parent[y] = x
rank[x] += 1
def KruskalMST(self):
result = []
i=0
e=0
self.graph = sorted(self.graph, key=lambda item: item[2])
 parent = []
rank = []
for node in range(self.V):
parent.append(node)
rank.append(0)
while e < self.V - 1:
u, v, w = self.graph[i]
i=i+1
x = self.find(parent, u)
y = self.find(parent, v)
if x != y:
e=e+1
result.append([u, v, w])
self.union(parent, rank, x, y)
minimumCost = 0
print("Edges in the constructed MST")
for u, v, weight in result:
minimumCost += weight
print("%d -- %d == %d" % (u, v, weight))
print("Minimum Spanning Tree", minimumCost)
if __name__ == '__main__':
g = Graph(4)
g.addEdge(0, 1, 10)
g.addEdge(0, 2, 6)
g.addEdge(0, 3, 5)
g.addEdge(1, 3, 15)
g.addEdge(2, 3, 4)
g.KruskalMST()

5.1.1. Output

5.2. Complexity analysis

The time complexity of Kruskal's algorithm is determined by the sorting of edges, which takes O(E * logE) time.
Following the sorting phase, the algorithm iterates through all edges, employing the find-union algorithm. The find
and union operations have a maximum time complexity of O(logV). Consequently, the overall time complexity is
expressed as O(E * logE + E * logV). Considering that the value of E can be at most O(V^2), the logarithmic terms
O(logV) and O(logE) are equivalent. Hence, the consolidated time complexity is presented as O(E * logE) or O(E *
logV).

5.3. Evaluation measures

Table. 3 highlights the execution time of Kruskal's algorithm across varying graph sizes. The algorithm
demonstrates consistent performance, indicating its scalability in constructing minimum spanning trees. The gradual
increase in execution time aligns with expectations, showcasing Kruskal's reliability for sparse graphs. These results
contribute valuable insights into the algorithm's efficiency in constructing optimal spanning trees.
Table 3

Execution time comparison

Number of vertices Number of edges Execution time(ms)

50 100 2.34
75 150 3.78
100 200 4.89
125 250 6.12
150 300 7.92
175 350 9.15
200 400 10.28

5.4. Observations
 The execution time of Kruskal's algorithm tends to increase gradually with the growth of the graph size. The
algorithm demonstrates scalability and efficiency in constructing minimum spanning trees for sparse graphs. Fig. 7
visually illustrates the algorithm's consistent performance across varying graph sizes.

EXECUTION TIME COMPARISON

400
300
200
100
0
1 2 3 4 5 6 7

Number of Vertices Number of Edges Execution Time (ms)

Fig. 7. Execution time vs Number of vertices

6. Conclusion

Kruskal's algorithm stands as a prominent example of a greedy approach for solving the minimum spanning tree
(MST) problem in graph theory. Throughout this discussion, we delved into the detailed mechanics of Kruskal's
algorithm, elucidating its step-by-step process of iteratively selecting edges with the minimum weight to construct an
optimal spanning tree. The algorithm's simplicity, efficiency, and scalability make it an invaluable tool in various
applications, such as network design, circuit layout, and urban planning. By providing a comprehensive explanation
and example, this discourse aimed to illuminate the algorithm's practical implementation. Moreover, the evaluation of
execution times across different vertices showcases its adaptability and efficiency in addressing real-world scenarios,
reaffirming Kruskal's algorithm as a versatile and impactful solution for minimum spanning tree problems in diverse
computational contexts.

References
[1] A. Katsigiannis, N. Anastopoulos, K. Nikas and N. Koziris, "An Approach to Parallelize Kruskal's Algorithm Using Helper Threads," 2012
IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum, Shanghai, China, 2012, pp. 1601-1610

[2] M. Aissa, A. B. Mnaouer and A. Belghith, "A novel extension of Kruskal's algorithm in multicast routing," 2009 IEEE 34th Conference on
Local Computer Networks, Zurich, Switzerland, 2009, pp. 289-292

[3] D. Rachmawati, Herriyance and F. Y. Putra Pakpahan, "Comparative Analysis of the Kruskal and Boruvka Algorithms in Solving Minimum
Spanning Tree on Complete Graph," 2020 International Conference on Data Science, Artificial Intelligence, and Business Analytics (DATABIA),
Medan, Indonesia, 2020, pp. 55-62

[4] F. Huang, P. Gao and Y. Wang, "Comparison of Prim and Kruskal on Shanghai and Shenzhen 300 Index Hierarchical Structure Tree," 2009
International Conference on Web Information Systems and Mining, Shanghai, China, 2009, pp. 237-241

[5] S. Tao, L. Xiuzhi, W. Yanyan, C. Yanli and C. Dayin, "A matlab simulation of the kruskal algorithm for erecting communication network,"
2011 International Conference on Electronics, Communications and Control (ICECC), Ningbo, China, 2011, pp. 41-43

[6] Wamiliana, M. Usman, D. Sakethi, R. Yuniarti and A. Cucus, "The hybrid of depth first search technique and Kruskal's algorithm for solving
the multiperiod degree constrained minimum spanning tree problem," 2015 4th International Conference on Interactive Digital Media (ICIDM),
Bandung, Indonesia, 2015, pp. 1-4

[7] T. D. Sudhakar and K. N. Srinivas, "Power system restoration based on Kruskal's algorithm," 2011 1st International Conference on Electrical
Energy Systems, Chennai, India, 2011, pp. 281-287

[8] P. B. Guttoski, M. S. Sunye and F. Silva, "Kruskal's Algorithm for Query Tree Optimization," 11th International Database Engineering and
Applications Symposium (IDEAS 2007), Banff, AB, Canada, 2007, pp. 296-302

[9] Qian Ren and Xinjian Zhuo, "Application of an improved K-means algorithm in gene expression data analysis," 2011 IEEE International
Conference on Systems Biology (ISB), Zhuhai, China, 2011, pp. 87-91
[10] U. B. Sayata and N. P. Desai, "An algorithm for Hierarchical Chinese postman problem using minimum spanning tree approach based on
Kruskal's algorithm," 2015 IEEE International Advance Computing Conference (IACC), Banglore, India, 2015, pp. 222-227

[11] P. O. Kalonda and A. M. Omekanda, "Kruskal's Algorithm, Vogel's Approximation and Modified Distribution Methods for the Design of
Optimal Electrical Networks in the Democratic Republic of Congo," 2020 IEEE PES/IAS PowerAfrica, Nairobi, Kenya, 2020, pp. 1-5

[12] M. Mosbah, S. Arif, R. D. Mohammedi and A. Hellal, "Optimum dynamic distribution network reconfiguration using minimum spanning
tree algorithm," 2017 5th International Conference on Electrical Engineering - Boumerdes (ICEE-B), Boumerdes, Algeria, 2017, pp. 1-6

[13] E. Çakır and Z. Ulukan, "An Intuitionistic Fuzzy MCDM Approach Adapted to Minimum Spanning Tree Algorithm for Spreading Content
on Social Media," 2021 IEEE 11th Annual Computing and Communication Workshop and Conference (CCWC), NV, USA, 2021, pp. 0174-0179

[14] R. Berger, S. Dubuisson and C. Gonzales, "Fast multiple histogram computation using Kruskal's algorithm," 2012 19th IEEE International
Conference on Image Processing, Orlando, FL, USA, 2012, pp. 2373-2376

[15] P. Ayegba, J. Ayoola, E. Asani and A. Okeyinka, "A Comparative Study Of Minimal Spanning Tree Algorithms," 2020 International
Conference in Mathematics, Computer Engineering and Computer Science (ICMCECS), Ayobo, Nigeria, 2020, pp. 1-4

[16] P. Malji and S. Sakhare, "Significance of entropy correlation coefficient over symmetric uncertainty on FAST clustering feature selection
algorithm," 2017 11th International Conference on Intelligent Systems and Control (ISCO), Coimbatore, India, 2017, pp. 457-463

[17] N. R. Latha, G. Shyamala and G. R. Prasad, "Exploring the parallel implementations of the three classical MST algorithms," 2017
International Conference on Inventive Communication and Computational Technologies (ICICCT), Coimbatore, India, 2017, pp. 340-346

[18] A. Shrestha, S. K. Jha, B. Shah and B. R. Gautam, "Optimal grid network for rural electrification of Upper Karnali Hydro Project affected
area," 2016 IEEE Region 10 Humanitarian Technology Conference (R10-HTC), Agra, India, 2016, pp. 1-5

[19] A. Gagarin, S. Hussain and L. T. Yang, "Distributed Search for Balanced Energy Consumption Spanning Trees in Wireless Sensor
Networks," 2009 International Conference on Advanced Information Networking and Applications Workshops, Bradford, UK, 2009, pp. 1037-
1042

[20] D. Kumar, M. Palaniswami, S. Rajasegarar, C. Leckie, J. C. Bezdek and T. C. Havens, "clusiVAT: A mixed visual/numerical clustering
algorithm for big data," 2013 IEEE International Conference on Big Data, Silicon Valley, CA, USA, 2013, pp. 112-117

[21] C. Deng and X. Chu, "Topology Estimation of Power Distribution Grids Using Graphical Model," 2020 IEEE/IAS Industrial and
Commercial Power System Asia (I&CPS Asia), Weihai, China, 2020, pp. 784-788

[22] C. Guada, D. Gómez, J. T. Rodríguez, J. Yáñez and J. Montero, "Fuzzy Image Segmentation Based on the Hierarchical Divide and Link
Clustering Algorithm," 2015 10th International Conference on Intelligent Systems and Knowledge Engineering (ISKE), Taipei, Taiwan, 2015, pp.
12-17

[23] D. P. Montoya and J. M. Ramirez, "A minimal spanning tree algorithm for distribution networks configuration," 2012 IEEE Power and
Energy Society General Meeting, San Diego, CA, USA, 2012, pp. 1-7

[24] N. Bisht and A. Basava, "Feature selection using Markov clustering and maximum spanning tree in high dimensional data," 2016 Ninth
International Conference on Contemporary Computing (IC3), Noida, India, 2016, pp. 1-6

[25] D. P. Montoya, J. M. Ramirez and J. R. Zuluaga, "Multi-objective optimization for reconfiguration and capacitor allocation in distribution
systems," 2014 North American Power Symposium (NAPS), Pullman, WA, USA, 2014, pp. 1-6

[26] W. N. Bhukya and A. Singh, "A study on energy issues in construction of all-to-all minimum power broadcast (A2A MPB) trees in wireless
networks," 2013 International Conference on Advances in Computing, Communications and Informatics (ICACCI), Mysore, India, 2013, pp.
1496-1501

[27] R. Feng et al., "Recommendation Pattern of Electricity Sales Package based on Improved Clustering and Spark Framework," 2022 IEEE
International Conference on Advances in Electrical Engineering and Computer Applications (AEECA), Dalian, China, 2022, pp. 293-299

[28] Y. Zuo, "Target Detection System of Agricultural Economic Output Efficiency Based on Kruskal Algorithm," 2022 IEEE 2nd International
Conference on Mobile Networks and Wireless Communications (ICMNWC), Tumkur, Karnataka, India, 2022, pp. 1-5

[29] M. S. Batta, Z. Aliouat, H. Mabed and M. Merah, "An Improved Lifetime Optimization Clustering using Kruskal’s MST and Batteries Aging
for IoT Networks," 2022 International Symposium on Networks, Computers and Communications (ISNCC), Shenzhen, China, 2022, pp. 1-6

[30] Z. Wang, G. Chen, L. Chen, D. Jia, M. Wu and T. Kang, "Fault Recovery Strategy of Distribution Network with Distributed Generation
Based on Island Division," 2022 4th International Conference on Smart Power & Internet Energy Systems (SPIES), Beijing, China, 2022, pp.
1217-1223

[31] Cormen, Thomas H., et al. "Introduction to Algorithms." MIT press, 2009. (Chapter 23, Minimum Spanning Trees)

[32] Kruskal, Joseph B. "On the shortest spanning subtree of a graph and the traveling salesman problem." Proceedings of the American
Mathematical society 7.1 (1956): 48-50.