Volume-3,Issue-5,May2025 International Journal of Modern Science and Research Technology

ISSN NO-2584-2706

Hybridization Approach for Ant Lion Optimizer

(ALO) using Particle Swarm Optimization (PSO)
Sandhya Dahake; Yogesh Sonvane
Ketki Sakharkar; Ruchika Bhadre
Department of Master in Computer Application, MCA
G H Raisoni College of Engineering and Management, Nagpur, Maharashtra, India.

Abstract which the Ant-lion Optimization Algorithm

Optimization algorithms are essential for (ALO) has gained attention for its strong
addressing intricate real-world challenges by optimization capabilities. However, despite its
effectively identifying optimal or near-optimal effectiveness, ALO can sometimes face
solutions. The Ant-lion Optimization limitations, such as slow convergence or
Algorithm (ALO) has proven to be effective in getting trapped in local optima. To overcome
various optimization scenarios; however, its these drawbacks, researchers have explored
performance can often be enhanced through hybridization techniques, where ALO is
hybridization with other meta-heuristic combined with other optimization methods to
methods. This research examines the effects of enhance its overall performance.
integrating Particle Swarm Optimization One promising hybrid approach involves
(PSO) with the ALO algorithm by contrasting integrating Particle Swarm Optimization
the results obtained from the original ALO (PSO) with ALO. PSO is well known for its
with those from the hybridized approach. The ability to quickly explore solution spaces and
efficacy of the hybrid method was assessed improve convergence speed, making it a
using a range of benchmark functions (F1 to suitable complement to ALO’s search
F23), which encompass various optimization mechanism. By combining these two
problems, including uni-modal, multi-modal, techniques, it is possible to leverage their
and fixed-point functions. The findings strengths while mitigating individual
indicate that the PSO-ALO hybridization weaknesses. This study examines the impact
significantly boosts optimization performance of PSO-ALO hybridization by comparing the
for several functions, such as F1, F3, F6, F7, results obtained using the original ALO
and F9, where the function values showed a algorithm with those achieved through the
notable decrease, reflecting improved solution hybrid method.
accuracy and convergence rates. This work proposed a novel nature-inspired
algorithm called ALO. The hunting behavior
Keywords: Hybridization, PSO, Nature- of ant-lions and entrapment of ants in ant-lions
Inspired Algorithms, Benchmark, Meta- traps were the main inspirations for this
heuristic Optimization. algorithm. Several operators were proposed
and mathematically modeled for equipping the
1. Introduction ALO algorithm with high exploration and
Optimization plays a crucial role in solving exploitation. After the survey it was observed
complex real-world problems by efficiently by comparing both original values and
identifying optimal or near-optimal solutions. hybridization values results was very
Various meta-heuristic algorithms have been impressive. An optimization-based method
developed to tackle such challenges, among was proposed in this paper for the problem of

IJMSRT25MAY067 www.ijmsrt.com 300

DOI: https://doi.org/10.5281/zenodo.15497012
Volume-3,Issue-5,May2025 International Journal of Modern Science and Research Technology
ISSN NO-2584-2706

instance reduction to obtain better results in 3. Literature Review

terms of many metrics in both balanced and Ants which consider as candidate solutions
imbalanced data. A new modified ant-lion move randomly in search of food which is
optimization (MALO) method was adapted for optimal solutions. Their movements were
this task after validating its ability in terms of modeled using random walk functions. On
optimization compared to state-of-the-art the other side there are ant-lions which are
optimizers using benchmark functions. The dig cone shaped pit in the sand. As soon as
results obtained at 500 and 1000 iterations for ant falls into the pits ant-lion throw sand to
twenty-three and thirteen benchmark trap them. Then the best ant-lion i.e.
functions, respectively, demonstrated that the solutions are continuously updated to
proposed MALO algorithm could escape the ensure that the algorithm finds optimal
local optima and provide a better convergence solution. After that once the ant falls into
rate as compared to the basic ALO algorithm the pit , the ant-lion pulls and consumes it
and state-of-the-art optimizers. i.e. it will keep repeating the solution till it
Additionally, instance reduction results from reaches the optimize
MALO were compared to basic ant-lion level means it gives proper solution of
Optimization and some well-known complex problems by repeating it multiple
optimization algorithms on 15 balanced and times until the best solution is found. As
imbalanced datasets to test the performance on this algorithm results was very impressive.
reducing instances of the training data. Fig.1. Optimization Algorithm Diagram
Furthermore, ant-lion optimization and
MALO were used to perform training data
reduction for 18 over-sampled imbalanced
datasets, and the
reduced datasets were classified by SVM in all
experiments. The results were also compared
with one novel resampling method.

2. Mathematical Expression of Pso

Table.1. Standard Benchmark Functions

Nature-inspired algorithms are widely used for

solving complex optimization problems in
engineering, computer science, bio-
informatics, and machine learning. They can
handle large search spaces and nonlinear
problems where traditional algorithms
struggle. Strong Balance between exploration
and exploitation for NIAs incorporate both
exploration (global search) and exploitation

IJMSRT25MAY067 www.ijmsrt.com 301

DOI: https://doi.org/10.5281/zenodo.15497012
Volume-3,Issue-5,May2025 International Journal of Modern Science and Research Technology
ISSN NO-2584-2706

(local refinement) mechanisms, allowing them 4. Result and Discussion

to escape local optima and high-quality Obtained results demonstrated that the
solutions. proposed MALO was superior in minimizing
Nature-inspired algorithms provide efficiency, the number of training set instances, hence
flexibility and adaptability for solving maximizing the classification performance
complex optimization problems. Their while reducing the run time compared to
biological and natural inspirations make them state-of-the-art methods used to reduce the
an exciting field for research with continuous original balanced and imbalanced datasets
advancements real-world problem-solving. without the need to perform oversampling
pre-processing methods which consume
Table.2. Algorithms and its Authors computational time and memory space.
MALO reduced the instances of over-
Algorithm Algorithm Author Name Year sampled imbalanced datasets with better
No. Name performance compared to the full over-
1. Hybrid Ant T. R. Mahesh, 2024
sampled data set and the recently proposed
Lion Mutated (senior member
Ant Colony ,IEEE), ACOR instance reduction method, ALO,
Optimizer D.Santhakumar, GWO, and WOA.
Technique A.Balajee , Fig.2. Results of Benc hmarks Functions
With Particle H.S.Shreenidhi
Swarm , V.Vinoth
Optimization Kumar ,
for Leukemia (Member,
Prediction IEEE), and
Using Jonnakuti
Microarray Rajkumar
Gene Data Annand
2. The Ant Lion S.Mirjalili 2015
Optimizer
3. A V. Pathak, 2022
comprehensive S.Gangwar,
survey on the Ramanpreet
ant lion Singh,
optimizer, A.Srivastava,
variants and Mithilesh K.
applications Dikshit
4. A genetic M.Rojas, 2023
operators- A.Olivera,
based Ant P.Vidal
Lion
Optimizer for
training a
medical multi-
layer
perceptron
5. Optimal real S.Verma, 2016
power V.Mukherjee
rescheduling
of generators
for congestion
management
using a novel
ant lion
optimizer

IJMSRT25MAY067 www.ijmsrt.com 302

DOI: https://doi.org/10.5281/zenodo.15497012
Volume-3,Issue-5,May2025 International Journal of Modern Science and Research Technology
ISSN NO-2584-2706

IJMSRT25MAY067 www.ijmsrt.com 303

DOI: https://doi.org/10.5281/zenodo.15497012
Volume-3,Issue-5,May2025 International Journal of Modern Science and Research Technology
ISSN NO-2584-2706

5. Conclusion
The performance of this hybrid approach was [2] R. Eberhart and J. Kennedy, ‘‘A new
tested across 23 benchmark functions (F1 to optimizer using particle swarm theory,’’ in
F23), which cover different types of Proc. 6th Int. Symp. Micro Mach. Hum. Sci.
optimization challenges, including uni-modal, (MHS), pp. 39–43, 1995.
multi-modal, and fixed-point problems. The
results indicate that for certain functions, such [3] M. Dorigo and M. Birattari, ‘‘Ant colony
as F1, F3, F6, F7, and F9, the hybrid optimization,’’ in Encyclopedia of Machine
algorithm significantly improved accuracy and Learning. New York, NY, USA: Springer, pp.
convergence speed by reducing function 36–39, 2011.
values. Additionally, function F8 showed
better minimization performance after [4] D. Karaboga, ‘‘An idea based on honey
hybridization. bee swarm for numerical optimization,’’ Dept.
However, the hybrid approach did not lead to Comput. Eng., Erciyes Univ., Kayseri,
improvements in all cases. In functions like Turkey, Tech. Rep. tr06, 2005.
F2, F4, F5, F10, F11, F13, and F22, the
function values increased, indicating a decline [5] D. Karaboga and B. Basturk, ‘‘A powerful
in performance compared to the standalone and efficient algorithm for numerical function
ALO method. Moreover, some functions, such optimization: Artificial bee colony (ABC)
as F14, F16, F17, F18, F19, and F21, algorithm,’’ J. Global Optim., vol. 39, no. 3,
remained unchanged, suggesting that ALO pp. 459–471, Oct. 2007.
had already reached optimal or near-optimal
solutions, making hybridization unnecessary. [6] S.Kirkpatrick, C. D.Gelatt, and M.
These findings highlight that while PSO-ALO P.Vecchi, ‘‘Optimization by simulated
hybridization can enhance optimization annealing,’’ Science, vol. 220, no. 4598, pp.
performance for many problems, it is not 671–680, 1983.
universally superior to standalone. The
effectiveness of hybridization depends on the [7] A. H.Gandomi and A. H. Alavi, ‘‘Krill
characteristics of the function being herd: A new bio-inspired optimization
optimized. Therefore, future research should algorithm,’’ Commun. Nonlinear Sci. Numer.
focus on developing adaptive hybridization Simul., vol. 17, no. 12, pp. 4831–4845, Dec.
techniques that can dynamically adjust the 2012.
influence of PSO and ALO based on the
nature of the problem. Such approaches could [8] X.-S.Yang and S.Deb, ‘‘Cuckoo search
help achieve more consistent improvements via Lévy flights,’’ in Proc. World Congr.
across a wider range of optimization Nature Biologically Inspired Comput.
challenges. (NaBIC), 2009, pp. 210–214.

6. References [9] Z. W. Geem, J. H. Kim, and G. V.

[1] X.-S. Yang and M. Karamanoglu, Loganathan, ‘‘A new heuristic optimization
‘‘Swarm intelligence and bio-inspired algorithm: Harmony search,’’ Simulation, vol.
computation: An overview,’’ in Swarm 76, no. 2, pp. 60–68, 2001.
Intelligence and Bio-Inspired Computation.
Amsterdam, The Netherlands: Elsevier, pp. 3–
23, 2013.

IJMSRT25MAY067 www.ijmsrt.com 304

DOI: https://doi.org/10.5281/zenodo.15497012
Volume-3,Issue-5,May2025 International Journal of Modern Science and Research Technology
ISSN NO-2584-2706

[10] X.-S.Yang, Nature-Inspired Meta- Artificial Intelligence. Cambridge, MA, USA:

heuristic Algorithms. Frome, U.K.: Luniver MIT Press, 1992.
Press, 2010.
[16] F. Glover, ‘‘Heuristics for integer
[11] X. Meng, Y. Liu, X. Gao, and H. Zhang, programming using surrogate constraints,’’
‘‘A new bio-inspired algorithm: Chicken Decis. Sci., vol. 8, no. 1, pp. 156–166, Jan.
swarm optimization,’’ in Proc. Int. Conf. 1977.
Swarm Intell. Cham, Switzerland: Springer, [17] F. Glover and C. Mcmillan, ‘‘The
2014, pp. 86–94. general employee scheduling problem. An
integration of MS and AI,’’ Comput. Oper.
[12] S. Mirjalili, S. M. Mirjalili, and A. Res., vol. 13, no. 5, pp. 563–573, Jan. 1986.
Lewis, ‘‘Grey wolf optimizer,’’ Adv.Eng.
Softw., vol. 69, pp. 46–61, Mar. 2014. [18] J. D. Farmer, N. H. Packard, and A. S.
Perelson, ‘‘The immune system, adaptation,
[13] A.G.Hussien, A.E.Hassanien, and machine learning,’’ Phys. D, Nonlinear
E.H.Houssein, M.Amin, and A.T.Azar,‘‘New Phenomena, vol. 22, nos. 1–3, pp. 187–204,
binary whale optimization algorithm for Oct. 1986.
discrete optimization problems,’’ Eng. Optim.,
pp. 1–15, Jun. 2019. [19] P. Moscato, ‘‘On evolution, search,
optimization, genetic algorithms and martial
[14] S. Mirjalili, ‘‘The ant lion optimizer,’’ arts: Towards memetic algorithms,’’ Caltech
Adv. Eng. Softw., vol. 83, pp. 80–98, May Concurrent Comput. Program, Monterey, CA,
2015. USA, C3P Rep. 826, 1989.

[15] J. H. Holland, Adaptation in Natural and [20] R. Y. Rubinstein, ‘‘Optimization of

Artificial Systems: An Introductory Analysis computer simulation models with rare
with Applications to Biology, Control, and events,’’ Eur. J. Oper. Res., vol. 99, no. 1, pp.
89–112, 1997.

IJMSRT25MAY067 www.ijmsrt.com 305

DOI: https://doi.org/10.5281/zenodo.15497012