Volume-3-Issue-6-June,2025 International Journal of Modern and Research Technology

ISSN NO-2584-2706

An Improved Hummingbird Algorithm by

Hybridization for Complex Optimization
Anshula Thosar; Rohit Kaikade
Mayuri Rangari; Prathmesh Kumar Dubey
Department of Master in Computer Applications, GHRCEM, India,

Abstract: The proposed algorithm tried to improve

The proposed algorithm is an improved the outcomes. The original Artificial
algorithm by hybridization for complex Hummingbird Algorithm was tested by
optimization. Algorithm will be tested by using hybridization technique with
using hybridization technique with Artificial Simulated Annealing Algorithm focused to
Hummingbird Algorithm (AHA) and enhance better values. Here, among
Simulated Annealing (SA), focuses to give techniques of improving the AHA
better solution. Here, the 23 functions will be algorithm’s results, the most suitable
applied and tested to compare the hybridized hybridization technique was used for
algorithm with existing algorithm. After obtaining better results. Here, the 23
testing, better results will be found in benchmark functions were tested to
hybridized algorithm using standard compare the hybridized algorithm with
benchmark functions. existing algorithm. After testing, better
results were found in 14 functions.
Keywords:
Algorithm, Functions, Optimization, 2. Proposed Optimization Algorithm
Hybridization, AHA-SA. The purpose of the Artificial Hummingbird
Algorithm (AHA) was inspired by the
1. Introduction special flight skills and intelligent foraging
The Artificial Hummingbird Algorithm strategies of hummingbirds in nature. Three
(AHA) was inspired by the special flight skills foraging strategies of hummingbirds,
and intelligent foraging strategies of including the guided foraging, territorial
hummingbirds in nature. Three foraging foraging, and migrating foraging, were
strategies of hummingbirds, including the implemented. . The motive of choosing this
guided foraging, territorial foraging, and algorithm was its results were found great.
migrating foraging, were implemented. The hybridization of AHA with HA seeks
Moreover, three kinds of flight skills utilized to combine strong optimization ability and
in the foraging strategies such as the axial, robust mechanism.
diagonal, and omnidirectional flights, are The nature inspired algorithms were
modelled. Specially, a visit table mimicking differentiated into four types which are
the supernormal memory ability of Physics-based, Human behavior-based,
hummingbirds was constructed to guide the Evolution-based and Swarm based. These
hummingbirds in the algorithm for performing algorithms use Physics-based techniques,
the global optimization.[6] Human-related techniques, Evolutionary
techniques and Swarm intelligence
techniques respectively.

IJMSRT25JUN25 www.ijmsrt.com 142

DOI: https://doi.org/10.5281/zenodo.15614740
Volume-3-Issue-6-June,2025 International Journal of Modern and Research Technology
ISSN NO-2584-2706

2.1. Classification of Algorithms

Sr. Algorithm Author

No. Year
Name Name
James
Particle Swarm
1. Kennedy et 1995
1.1.Algorithms & Authors Optimization
al
Xin-She
2. Bat Algorithm 2010
Table 1: Algorithms and Authors [6] Yang
Genetic John R.
3. Programming Koza 1992
2.2. Steps Biogeography-
1. The original Artificial Hummingbird 4. Based Dan Simon 2008
Algorithm (AHA) was tested using 23 Optimization
standard functions. Brain Storm
5. Optimization Yuhui Shi 2011
2. The original algorithm (AHA) was
Teaching-
hybridized with another algorithm (SA) for Learning- R.V. Rao et
obtaining best values. 6. Based 2011
al
3. Iterations were carried out for each Optimization
function. 7.
Harmony Zong Woo
2001
4. Obtained values for another algorithm (SA) Search Geem et al
River Xavier
using 23 functions.
8. Formation Sánchez et 2007
5. Compared the best optimal value found by Dynamics al
AHA and the value after hybridization with
the SA.
6. Results were found good in 14 benchmark
functions.

3. Functions & Equations

Table 2: Standard UM Benchmark functions

[6]

IJMSRT25JUN25 www.ijmsrt.com 143

DOI: https://doi.org/10.5281/zenodo.15614740
Volume-3-Issue-6-June,2025 International Journal of Modern and Research Technology
ISSN NO-2584-2706

4. RESULTS & DISCUSSION The best optimal value by function 5

 Function 1: after hybridization was found to be
2.07E+03.

 Function 6:

The best optimal value by function 1 after

hybridization was found to be 1.95E+03.
The best optimal value by function 6 after
 Function 2: hybridization was found to be 1.73E+03.

 Function 7:

The best optimal value by function 2 after

hybridization was found to be 1.88E+03.

 Function 3: The best optimal value by function 7 after

hybridization was found to be 1.78E+03.

 Function 8:

The best optimal value by function 3 after

hybridization was found to be 1.91E+03.

 Function 4:
The best optimal value by function 8 after
hybridization was found to be 1.92E+03.

 Function 9:

The best optimal value by function 4 after

hybridization was found to be 2.14E+03.

 Function 5:
The best optimal value by function 9 after
hybridization was found to be 1.81E+03.
 Function 10:

IJMSRT25JUN25 www.ijmsrt.com 144

DOI: https://doi.org/10.5281/zenodo.15614740
Volume-3-Issue-6-June,2025 International Journal of Modern and Research Technology
ISSN NO-2584-2706

The best optimal value by function 10 after  Function 16:

hybridization was found to be 1.94E+03.
 Function 11:

The best optimal value by function 16 after

hybridization was found to be 2.01E+03.
The best optimal value by function 11 after  Function 17:
hybridization was found to be 2.05E+03.

 Function 12:

The best optimal value by function 17 after

hybridization was found to be 1.97E+03.
 Function 18:
The best optimal value by function 12 after
hybridization was found to be 2.00E+03.

 Function 13:

The best optimal value by function 18 after

hybridization was found to be 1.97E+03.
 Function 19:

The best optimal value by function 13 after

hybridization was found to be 1.80E+03.

 Function 14:
The best optimal value by function 19 after
hybridization was found to be 1.73E+03.
 Function 20:

The best optimal value by function 14 after

hybridization was found to be 1.97E+03.
The best optimal value by function 20 after
 Function 15: hybridization was found to be 1.89E+03.
 Function 21:

The best optimal value by function 15 after

hybridization was found to be 1.66E+03. The best optimal value by function 21 after
hybridization was found to be 2.26E+03.

IJMSRT25JUN25 www.ijmsrt.com 145

DOI: https://doi.org/10.5281/zenodo.15614740
Volume-3-Issue-6-June,2025 International Journal of Modern and Research Technology
ISSN NO-2584-2706

5. CONCLUSION
 Function 22: The hybridized Artificial Hummingbird
Algorithm with Simulated Annealing
Algorithm resulted in superior
performance in convergence speed,
solution quality, and robustness. The
hybridized algorithm of AHA with SA
The best optimal value by function 22 after
was tested with 23 benchmark functions
hybridization was found to be 1.66E+03.
and in 14 functions best values were
found considered as good results.
 Function 23:
6. REFERENCES
[1] W. Y. Lin, “A novel 3D fruit fly
optimization algorithm and its applications
The best optimal value by function 23 after in economics,” Neural Comput. Appl.,
hybridization was found to be 1.75E+03. 2016, doi: 10.1007/s00521-015-1942-8.
Benchmark (AHA) Algorithm (AHA+SA)
Algorithm
[2] M. Mafarja et al., “Evolutionary
Functions Values
Values Population Dynamics and Grasshopper
F1 6.83E-303 1.95E+03 Optimization approaches for feature
F2 2.23E-162 1.88E+03 selection problems,” Knowledge-Based
F3 1.44E-295 1.91E+03 Syst., vol. 145, pp. 25–45, 2018, doi:
F4 2.55E-138 2.14E+03 10.1016/j.knosys.2017.12.037.
F5 2.53E+01 2.07E+03
F6 0 1.73E+03 [3] E.-S. M. El-Kenawy, M. M. Eid, M.
F7 2.74E-06 1.78E+03 Saber, and A. Ibrahim, “MbGWO-SFS:
F8 -1.21E+04 1.92E+03 Modified Binary Grey Wolf Optimizer
F9 0.00E+00 1.81E+03 Based on Stochastic Fractal Search for
F10 4.44E-16 1.94E+03
Feature Selection,” IEEE Access, 2020,
F11 0.00E+00 2.05E+03
doi: 10.1109/access.2020.3001151.
F12 2.19E-07 2.00E+03
F13 4.77E-01 1.80E+03
F14 0.998 1.97E+03
[4] Y. Cheng, S. Zhao, B. Cheng, S. Hou,
F15 0.00030749 1.66E+03 Y. Shi, and J. Chen, “Modeling and
F16 -1.0316 2.01E+03 optimization for collaborative business
F17 0.39789 1.97E+03 process towards IoT applications,” Mob.
F18 3 1.97E+03 Inf. Syst., 2018, doi:
F19 -3.8628 1.73E+03 10.1155/2018/9174568.
F20 -3.322 1.89E+03
F21 -10.1532 2.26E+03 [5] I. E. Grossmann, Global Optimization in
F22 -10.4029 1.66E+03 Engineering Design (Nonconvex
F23 -10.5364 1.75E+03 Optimization and Its Applications), vol. 9.
1996.
Table 3: Results & Discussion
R. V. Rao and G. G. Waghmare, “A new
[6]
optimization algorithm for solving complex
constrained design optimization problems,”
vol. 0273, no. April, 2016, doi:
10.1080/0305215X.2016.1164855

IJMSRT25JUN25 www.ijmsrt.com 146

DOI: https://doi.org/10.5281/zenodo.15614740
Volume-3-Issue-6-June,2025 International Journal of Modern and Research Technology
ISSN NO-2584-2706

technique and meteorological records,”

[7]X. Wang, T. M. Choi, H. Liu, and X. Yue, AIMS Energy, 2017, doi:
“A novel hybrid ant colony optimization 10.3934/energy.2017.5.798.
algorithm for emergency transportation
problems during post-disaster scenarios,” IEEE [12] R. Al-Hajj, A. Assi, and F. Batch, “An
Trans. Syst. Man, Cybern. Syst., 2018, doi: evolutionary computing approach for
10.1109/TSMC.2016.2606440. estimating global solar radiation,” in 2016
IEEE International Conference on Renewable
[8]M. Nouiri, A. Bekrar, A. Jemai, S. Niar, Energy Research and Applications, ICRERA
and A. C. Ammari, “An effective and 2016, 2017. doi:
distributed particle swarm optimization 10.1109/ICRERA.2016.7884553.
algorithm for flexible job-shop scheduling
problem,” J. Intell. Manuf., 2018, doi: [13] R. A. Meyers, “Classical and Nonclassical
10.1007/s10845-015-1039-3. Optimization Methods Classical and
Nonclassical Optimization Methods 1
[9]Y. Li, J. Wang, D. Zhao, G. Li, and C. Chen, Introduction 1 1.1 Local and Global
“A two-stage approach for combined heat and Optimality 2 1.2 Problem Types 2 1.3
power economic emission dispatch: Combining Example Problem: Fitting Laser-induced
multi-objective optimization with integrated Fluorescence Spectra 3 1.4 Criteria for
decision making,” Energy, 2018, doi: Optimization 4 1.5 Multicriteria Optimization
10.1016/j.energy.2018.07.200. 4,” Encycl. Anal. Chem., pp. 9678–9689,
2000, [Online].
[10]D. Yousri, T. S. Babu, and A. Fathy, “Recent [14] N. Steffan and G. T. Heydt,
methodology based Harris hawks optimizer for “Quadratic programming and related
designing load frequency control incorporated in techniques for the calculation of
multi-interconnected renewable energy plants,” locational marginal prices in
Sustain. Energy, Grids Networks, 2020, doi: distribution systems,” in 2012 North
10.1016/j.segan.2020.100352. American Power Symposium
(NAPS), 2012, pp. 1–6. doi:
[11] R. Al-Hajj and A. Assi, “Estimating solar 10.1109/NAPS.2012.6336310.
irradiance using genetic programming

IJMSRT25JUN25 www.ijmsrt.com 147

DOI: https://doi.org/10.5281/zenodo.15614740