Comparison of Ant Colony Optimization, Genetic

Optimization, and Hamming Distance Algorithms for Vector

Reduction

Kumar K K

, Suhas Shirol

, Kruthik Gupta B

, Ashwini M Pattennavar

, Spoorti Patil

Saroja V S

, Vijay H M

and Rajeshwari M

KLE Tech, KLE Technological University, Vidyanagar, Hubli, India

Electronics and Communication, KLE Technological University, Hubli, India

Keywords: Ant Colony Algorithm, Genetic Algorithm, Hamming Distance Algorithm, Test Vectors, K-Means, Verilog

HDL Implementation, Minimum Power Leakage.

Abstract: In this paper, we present the Ant colony algorithm, Genetic algorithm, and Hamming Distance algorithm for

vector reduction and reduced power leakage using algorithms. The main motivation is to compare Ant Colony

Optimization (ACO), Genetic Optimization (GO), and Hamming Distance Algorithms for vector reduction to

identify the most effective method for optimizing vector representation in various computational applications.

This step aims to reduce the number of vectors and reduce power leakage. This paper proposes a new approach

to the Ant colony algorithm based on switching activity, Hamming distance based on K-means clustering,

and a Genetic algorithm based on fitness for vector reduction. In the comparative analysis of power leakage

between algorithms, the Ant Colony algorithm exhibited the lowest power leakage of 0.17%. Therefore the

Ant Colony algorithm is the most efficient in terms of power consumption

1 INTRODUCTION

The rapid advancement of technology into the

nanometer scale has led to a significant increase in

sub-threshold leakage currents, which rise

exponentially as the supply voltage (Vdd) and

threshold voltage (Vth) decrease. In contemporary

CMOS technologies, sub-threshold leakage current

has become the dominant component of total leakage

current. Minimizing leakage power is especially

crucial for portable devices, which often operate in

standby mode, to extend battery life. Among various

techniques for reducing leakage power, Input Vector

Control (IVC) stands out due to its independence

from process technology parameters and its reliance

on the transistor stacking effect. IVC positions a

circuit in its minimum leakage state without

compromising performance.

Genetic Optimization, rooted in natural selection

and genetics principles, is an evolutionary algorithm

https://orcid.org/0009-0008-3924-2775

https://orcid.org/0009-0003-2257-3723

that iterates through generations of candidate

solutions to find the best fit. This process involves

selection, crossover, and mutation operations to

evolve a population of potential solutions over

successive generations. GO is particularly effective in

global search optimization problems and can

adaptively discover optimal subsets of vectors,

making it a powerful tool for reducing the

dimensionality of large datasets while preserving

essential information.

Ant Colony Optimization, inspired by the

foraging behavior of real ants, is a probabilistic

technique primarily used for solving combinatorial

optimization problems. Introduced by Marco Dorigo

in the early 1990s, ACO leverages a population of

artificial ants that iteratively construct solutions and

exchange information via pheromone trails. These

pheromones influence the likelihood of future ants

selecting specific paths, guiding the algorithm toward

optimal or near-optimal solutions. In vector

608

K K, K., Shirol, S., Gupta B, K., M Pattennavar, A., Patil, S., V S, S., H M, V. and M, R.

Comparison of Ant Colony Optimization, Genetic Optimization, and Hamming Distance Algorithms for Vector Reduction.

DOI: 10.5220/0013582500004664

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 3rd International Conference on Futuristic Technology (INCOFT 2025) - Volume 1, pages 608-614

ISBN: 978-989-758-763-4

reduction, ACO can efficiently identify relevant

components by exploring and exploiting the search

space through cooperative behavior.

Hamming Distance, a dissimilarity measure

between two strings of equal length, has been widely

used in error detection and correction codes. When

applied to vector reduction, algorithms based on

Hamming Distance focus on identifying and

removing redundant or less significant vectors by

comparing binary data representations. This approach

is particularly advantageous in scenarios where

binary or categorical data needs to be simplified

without losing critical information.

2 LITERATURE SURVEY

Several researchers have employed genetic

algorithms (GAs) to find the MLV. Chen et al.

introduced a genetic algorithm that uses an accurate

leakage current model to search for the MLV. This

approach demonstrated significant improvements in

leakage power reduction compared to traditional

methods. Xiaoying Zhao et al. enhanced this

approach by using Circuit Status Difference (CSD) as

the fitness function, which addressed some

limitations of the previous technique. Their method

showed that GAs could efficiently identify the MLV

with fewer iterations and reduced runtime compared

to random search methods. Implementing genetic

algorithms in hardware description languages

(HDLs) like Verilog has shown promising results

regarding convergence speed and computational

efficiency. The HDL-based approach allows for the

simulation and synthesis of test circuits, making

enforcing the MLV during standby mode easier to

reduce leakage power. (Thong, and, Champrasert,

2023), (Sun, 2023),(Zhou and Gao, 2024), (Zhang,

Li, et al. , 2023), (Pham, Nguyen, et al. , 2023).

The research paper explores Ant Colony

Optimization (ACO) and its application to the

Traveling Salesman Problem (TSP), inspired by the

foraging behavior of real ants to find the shortest path.

By simulating the collective intelligence of ant

colonies, ACO efficiently solves complex

optimization problems. The study focuses on

constructing graphs for TSP, associating components

with either edges or vertices and provides examples

with four cities to demonstrate ACO’s effectiveness.

ACO algorithms have proven adaptable and efficient

in finding near-optimal solutions across various

optimization scenarios. (Awad, Hawash, et al., 2023),

(Jayakumar and Khatri, 2007), (Lin, Qu, et al. , 2006),

(Chen, , et al. , 1998), (Leelarani, Madhavilatha, et al.,

2015)

The research paper presents a clustering algorithm

tailored for categorical data, utilizing the Hamming

distance as its metric. It transforms data into binary

form before clustering, demonstrating promising

performance in experiments on UCI machine learning

repository datasets. The results suggest the

algorithm's potential effectiveness across different

data types. The authors developed a clustering

algorithm for categorical data using the Hamming

distance metric. They transformed the data into

binary form to apply the Hamming distance

effectively. Performance evaluations on UCI machine

learning repository datasets showed the algorithm's

effectiveness, surpassing existing methods in

clustering categorical data. (Lan, 2023), (Starzec,

Starzec, et al. , 2024), (Dorigo, Birattari, et al. , 2006),

(Raman, Kumar, et al. , 2024)

3 DESIGN METHODOLOGY

3.1 Flowchart of Ant colony Algorithm

Figure 1: Flowchart of Ant colony Algorithm

The flowchart in Figure 1. illustrates the

procedural steps involved in the Ant Colony

Optimization (ACO) algorithm, tailored for a specific

problem involving path construction. Initially, the

Comparison of Ant Colony Optimization, Genetic Optimization, and Hamming Distance Algorithms for Vector Reduction

609

process begins with an initialization phase where

parameters such as pheromone levels are set, and a

population of artificial ants is created. Each ant then

constructs a path, visiting nodes or cities until they

have covered half of them. The algorithm then checks

if the number of ants encountering each other at nodes

exceeds a predefined threshold (ν). If this condition is

met, the paths of these ants are combined, which

likely involves integrating their solutions to enhance

the overall quality. If the threshold is not reached, the

ants continue constructing their paths until all paths

are complete. Once all paths are constructed, the

pheromone levels are updated based on the quality of

the constructed paths, reinforcing successful routes.

This iterative process continues until a termination

condition is met, such as a maximum number of

iterations for convergence to a satisfactory solution,

at which point the algorithm ends

3.2 Flowchart of Hamming Distance

Algorithm

Figure 2: Flowchart of Hamming Distance Algorithm

Figure 2. shows that the Hamming Distance starts

by taking two input vectors and checking if they have

the same length; if not, calculating the Hamming

distance is not possible. If the lengths are equal, each

corresponding element of the vectors is compared. If

the elements are the same, the distance for that

position is 0; if it is different, the distance is 1. This

process is repeated for all elements, and the final

Hamming distance is the sum of these individual

distances.

3.3 Flowchart of Genetic Algorithm

Figure 3: Flowchart of Genetic Algorithm

Figure 3. shows the genetic algorithm starts by

generating a random initial set of solutions called

chromosomes. Each chromosome's fitness is assessed

to determine how well it solves the problem. Two

chromosomes are then selected based on their fitness,

using methods like a roulette wheel or tournament

selection. These selected chromosomes undergo

crossover to produce offspring, combining parts of

the parents' chromosomes. A mutation is applied to

the offspring to introduce random changes and

maintain genetic diversity. The algorithm checks if

the stopping criterion is met; if not, it repeats the

evaluation and selection process with the new

population. Once the stopping criterion is met, the

algorithm outputs the best solution found and

terminates.

3.4 Final Design

The functional block diagram outlines a systematic

approach to optimize circuit design and reduce power

consumption is shown in Figure 4. It commences with

Test Vectors, initial input data intended for circuit

testing, which are processed initially in MATLAB to

prepare them for optimization. Following this

preparation, three distinct algorithms are applied. The

outputs from these algorithms are the Reduced

Vectors, subsequently integrated into the design flow

using the Cadence tool Within Cadence, Verilog

(HDL) code is generated from the Reduced Vectors

to instantiate and simulate the OR Gate Circuit, where

the actual hardware design and implementation occur.

INCOFT 2025 - International Conference on Futuristic Technology

610

Power measurement assesses the power consumption

of the implemented circuit, followed by power

analysis, which compares power metrics before and

after algorithmic optimizations. The resulting Power

Reduction quantifies the efficiency gains achieved

through algorithms compared to the initial Power

without Algorithm. Ultimately, Power with

Algorithms signifies the reduced power consumption

realized through systematic optimization techniques

in Verilog-based circuit design using MATLAB and

the Cadence tool.

Figure 4: Circuit Design for Optimizing Power Efficiency

4 RESULTS AND ANALYSIS

4.1 Results obtained in MATLAB

4.1.1 Ant Colony Algorithm

This Algorithm uses an ant colony algorithm to

reduce vectors based on the switching activity

between the test vectors. This algorithm ensures that

the sequence provided by the algorithm is more

efficient for testing the vectors, as shown in Figure 5

Figure 5: Results of Ant colony based on switching

4.1.2 Hamming Distance Algorithm

The k-means algorithm begins by randomly selecting

four initial centroids from the dataset. Each vector is

assigned to the nearest centroid based on the

Hamming distance, forming clusters. The centroids

are updated by calculating the mean for each bit

position of the vectors within each cluster. These

steps are repeated until the centroids stabilize,

indicating convergence, and after reduction, the

reduced vectors are given as input to any circuit as a

testbench, as shown in Figures 6 and 7

Figure 6: Results of Indices Corresponding Clusters and

Minimum Hamming Distance

Figure 7: The Centroid Values of each vector

4.1.3 Genetic Algorithm

A genetic algorithm is a computational optimization

technique that mimics the process of natural selection

to evolve a population of potential solutions to a

problem, iteratively improving them through

Comparison of Ant Colony Optimization, Genetic Optimization, and Hamming Distance Algorithms for Vector Reduction

611

selection, crossover, and mutation operations. In the

Genetic algorithm, we are using tournament selection

to reduce the size of vectors in the genetic algorithm.

Figure 8: Results of reduced vectors of Genetic algorithm

4.2 Results Analysis pf power

An analysis of power leakage reduction using three

different algorithms those are Ant Colony, Genetic,

and Hamming Distance is shown in Table 1. When no

algorithm is applied, the power leakage is 0.20 %

across the board. However, applying the Ant Colony

algorithm reduces power leakage to 0.17 %, showing

the most significant improvement. The Genetic and

Hamming Distance algorithms both reduce power

leakage to 0.19 %, which, while better than the

baseline, are not as effective as the Ant Colony

algorithm. Overall, all three algorithms help decrease

power leakage compared to not using any algorithm

Table 1: Comparison of power leakage between

algorithms in percentage

SL.N

ALGORITH

Power

leakage (with

Algorithm)

(%)

Power

leakage

(without

Algorithm

)

(

)

ANT

COLONY

0.17 0.20

GENETIC 0.19 0.20

HAMMING

DISTANCE

0.19 0.20

4.2.1 Power with Normal Test Vectors

Figure 9: Power with Normal Test Vectors

4.2.2 Power with Reduced Test Vectors of

ant colony algorithm

Figure 10: Power with reduced test vectors of ant colony

algorithm

4.2.3 Power with Reduced Test Vectors of

Hamming Distance Algorithm

Figure 11: Power Analysis for the reduced vectors of

Hamming Distance Algorithm

INCOFT 2025 - International Conference on Futuristic Technology

612

4.2.4 Power with Reduced Test Vectors of

Hamming Distance Algorithm

Figure 12: Power Analysis for the reduced vectors of

Genetic Algorithm

Figure 9,10,11 and 12 shows the detailed power

analysis for all the algorithms that were implemented

5 CONCLUSION

A comparative analysis of power leakage between

algorithms is conducted, and the Ant Colony

Optimization (ACO) algorithm exhibits the lowest

power leakage among other algorithms at 0.17 %.

This significant finding suggests that the Ant Colony

algorithm is the most efficient in power consumption,

making it an attractive option for applications where

energy efficiency is paramount. Furthermore, the

results demonstrate that using these algorithms can

lead to a notable reduction in power leakage, with the

power leakage without using the algorithm

consistently higher at 0.20 % across all three

algorithms. Therefore, the Ant Colony algorithm is

the top choice for minimizing power leakage, closely

followed by the Genetic and Hamming Distance

algorithms.

REFERENCES

S. Thong-ia and P. Champrasert, "Gene-Ants: Ant Colony

Optimization with Genetic Algorithm for Traveling

Salesman Problem Solving," 2023 International

Technical Conference on Circuits/Systems, Computers,

and Communications (ITC-CSCC), Jeju, Korea,

Republic of, 2023, pp. 1-5, doi: 10.1109/ITC-

CSCC58803.2023.10212945.

L. Sun, "Path Planning of Mobile Robot Based on Improved

Ant Colony Algorithm," 2023 IEEE 11th Joint

International Information Technology and Artificial

Intelligence Conference (ITAIC), Chongqing, China,

2023, pp. 985-989, doi:

10.1109/ITAIC58329.2023.10409046.

Y. Zhou and Z. Gao, "Simulation of Industrial Internet

Reliability Assessment Model Based on Ant Colony

Optimization Algorithm and Machine Learning," 2024

Asia-Pacific Conference on Software Engineering,

Social Network Analysis and Intelligent Computing

(SSAIC), New Delhi, India, 2024, pp. 248-252, doi:

10.1109/SSAIC61213.2024.00053.

S. Zhang, J. Li and Y. Li, "Reachable Distance Function for

KNN Classification," in IEEE Transactions on

Knowledge and Data Engineering, vol. 35, no. 7, pp.

7382-7396, 1 July 2023, doi:

10.1109/TKDE.2022.3185149

T. Xuan Pham, T. Tan Nguyen and H. Lee, "Hamming-

Distance Trellis Min-Max-Based Architecture for Non-

Binary LDPC Decoder," in IEEE Transactions on

Circuits and Systems II: Express Briefs, vol. 70, no. 7,

pp. 2390-2394, July 2023, doi:

10.1109/TCSII.2023.3241112.

A. Awad, A. Hawash and B. Abdalhaq, "A Genetic

Algorithm (GA) and Swarm-Based Binary Decision

Diagram (BDD) Reordering Optimizer Reinforced

With Recent Operators," in IEEE Transactions on

Evolutionary Computation, vol. 27, no. 3, pp. 535-549,

June 2023, doi: 10.1109/TEVC.2022.3170212.

Jayakumar, Nikhil, and Sunil P. Khatri. "An algorithm to

minimize leakage through simultaneous input vector

control and circuit modification." 2007 Design,

Automation & Test in Europe Conference & Exhibition.

IEEE, 2007.

Yuan, Lin, and Gang Qu. "A combined gate replacement

and input vector control approach for leakage current

reduction." IEEE transactions on very large scale

integration (vlsi) systems 14.2 (2006): 173-182.

Chen, Zhanping, et al. "Estimation of standby leakage

power in CMOS circuits considering accurate modeling

of transistor stacks." Proceedings of the 1998

international symposium on Low power electronics and

design. 1998.

Leelarani, V., and M. Madhavilatha. "Verilog

implementation of genetic algorithm for minimum

leakage vector in input vector control approach." 2015

International Conference on Signal Processing and

Communication Engineering Systems. IEEE, 2015.

Y. Lan, "Binary-like Real Coding Genetic

Algorithm," 2023 International Conference on Pattern

Recognition, Machine Vision and Intelligent

Algorithms (PRMVIA), Beihai, China, 2023, pp. 98-

102, doi: 10.1109/PRMVIA58252.2023.00023.

Grażyna Starzec, Mateusz Starzec, Leszek Rutkowski,

Marek Kisiel-Dorohinicki, Aleksander Byrski, Ant

colony optimization using two-dimensional pheromone

for single-objective transport problems, Journal of

Computational Science, Volume 79, 2024,102308,

ISSN 1877-7503

M. Dorigo, M. Birattari and T. Stutzle, "Ant colony

optimization," in

IEEE Computational Intelligence

Magazine, vol. 1, no. 4, pp. 28-39, Nov. 2006, doi:

10.1109/MCI.2006.329691.

R. Raman, V. Kumar, B. G. Pillai, D. Rabadiya, S. Patre

and R. Meenakshi, "The Impact of Enhancing the k-

Comparison of Ant Colony Optimization, Genetic Optimization, and Hamming Distance Algorithms for Vector Reduction

613

Means Algorithm through Genetic Algorithm

Optimization on High Dimensional Data Clustering

Outcomes," 2024 International Conference on

Knowledge Engineering and Communication Systems

(ICKECS), Chikkaballapur, India, 2024, pp. 1-5, doi:

10.1109/ICKECS61492.2024.10617268.

INCOFT 2025 - International Conference on Futuristic Technology

614