IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Available online at: http://publikasi.mercubuana.ac.id/index.php/ijiem

IJIEM (Indonesian Journal of Industrial Engineering & Management)

ISSN (Print) : 2614-7327 ISSN (Online) : 2745-9063

Foraging Bee Optimization Algorithm

Ebun Phillip Fasina*, Babatunde Alade Sawyerr, Shuaibu Babangida Alkassim
Department of Computer Science, University of Lagos, Lagos, Nigeria

ARTICLE INFORMATION A B S T R A C T
Article history:
Honeybees feed on pollen and nectar from flowers. Nectar
Received: 3 May 2023 to meet their energy requirements and pollen for protein and
Revised: 2 June 2023 other vital nutrients. The act of searching for these flowers
Accepted: 4 June 2023
by honeybees is called foraging. The foraging behaviour of
Category: Research paper bees depends on the profitability of nectar and pollen
sources as well as the needs of the colony. This behaviour is
Keywords:
Swarm intelligence
modelled by the Foraging Bee Optimization Algorithm
Nature-inspired metaheuristics (FBA) as metaphor for optimization. After initialization, the
Bee-inspired optimization algorithm algorithm loops through three phases based on bees’
Numerical optimization foraging behaviour –work, withdraw, and waggle (3W).
Particle swarm optimization Flowers are initialized randomly in the problem space.
During the waggle phase, bees are recruited to flowers with
profitable nectar sources. In the work phase, new flowers are
DOI: 10.22441/ijiem.v4i2.20275 discovered and memorized by bees. In the withdraw phase
bees remove unprofitable flowers from collective memory
and recalibrate for recruitment. The proposed FBA is tested
on three unimodal and twelve multimodal benchmark
functions. The result is compared with two other state-of-
the-art swarm intelligence algorithms, Artificial Bee Colony
(ABC) and Particle Swarm Optimization (PSO). Analysis of
comparison results shows FBA to be highly competitive,
outperforming PSO on all benchmarks and matching ABC
in overall performance.
This is an open access article under the CC–BY-NC license.
*Corresponding Author
Ebun Fasina
E-mail: [email protected]

1. INTRODUCTION Swarm Intelligence Optimization algorithms or

The study of the behavior of social organisms metaheuristics and have been applied to solve a
as a swarm in and outside their colonies led to diverse range of problems (Fakhermand &
Swarm Intelligence (SI) (Eberhart, Shi, & Derakhshani, 2023; Tzanetos & Dounias, 2020;
Kennedy, 2001; Janaki & Geethalakshmi, 2022; Engelbrecht, 2007; Alizadehsani, et al., 2023;
Selvaraj & Choi, 2020). SI is a discipline in Altshuler, 2023; Shahzad, et al., 2023; Kumar,
computer science that mimics the intelligence Chatterjee, Payal, & Rathore, 2022; Cruz, Maia,
displayed by social organisms (Kaswan, & de Castro, 2021).
Dhatterwal, & Kumar, 2023; Schumann, 2020).
This intelligence can be self-learning, healing, Nature-Inspired algorithms find approximate
or optimizing. Researchers model and create solutions to optimization problems, the solution
algorithms based on this intelligence. These can be local or global optimum depending on
algorithms are classified as Nature-Inspired or the set of constraints the optimization problem

How to Cite: Fasina, E. P., Sawyerr, B. A., & Alkassim, S. B. (2023). Foraging Bee Optimization Algorithm. IJIEM (Indonesian 99
Journal of Industrial Engineering & Management), 4(2), 99-112. https://doi.org/10.22441/ijiem.v4i2.20275
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

is subjected to. An optimization problem and nectar from flowers, and conveyance of
requires an objective function that may be pollen and nectar to the colony. Bees in charge
constrained or unconstrained to be maximized of foraging are called foragers. Each forager
or minimized. Optimization algorithms invoke modulates its behaviour in relation to the
the objective function to determine the fitness profitability of the nectar source – the more
of a large and varied selection of solutions to profitable the source, the higher the intensity of
determine the best or near optimum. foraging activity around the source, the more
Optimization techniques are mostly applied to repetitive and dancing (or waggle) at the nest
minimize cost or error, maximize profit, and pointing to the source, and the lower the
find optimal designs for engineering problems probability of abandoning the source. Without
or provide optimal decisions for operational and comparing sources, bee individually calculate
management problems. the absolute profitability of a source. The
collective nectar and pollen source selection by
Various natured-inspired algorithm has been a colony of bees is decentralized; it is a process
proposed among which are the Particle Swarm of natural selection where foragers from more
Optimization (PSO) by Kennedy and Eberhart profitable nectar sources continue to visit these
(Kennedy J. and Eberhart, 1995) which is sources over a long period of time and
inspired by simulation studies of the social eventually recruit bees from less profitable
behavior found in schools of fish and flocks of sources. In a typical foraging season, bees
birds. Bee Colony Optimization (BCO) by collect roughly 20 – 30 kg of pollen and 125kg
(Teodorovic & Dell’orco, 2005), Bee nectar which translate to between 1,125,000 and
Algorithm (BA) by (Pham, et al., 2005), 4,000,000 visits to flowers.
Artificial Bee Colony by (Karaboga, 2005) are
all inspired the foraging behavior of bee II. Bee Colonies as Metaphors for Swarm
colonies. Genetic recombination and natural Intelligence Algorithms
selection inspired the Generic Algorithm (GA) Agents in the Bee Algorithm (BA) first
proposing by (Holland, 1975). Studies of ant proposed by (Pham, et al., 2005) combined
colonies resulted in the Ant Colony randomized search of the problem space with
Optimization (ACO) algorithm by (Dorigo, neighborhood search in promising regions of
Colorni, & Maniezzo, 1991). Differential this space. BA is complex and can easily be
Evolution (DE) was proposed by (Storn & trapped in a local optimum. The Artificial Bee
Price, 1997), and Glowworm Swarm Colony (ABC) algorithm proposed by
Optimization was proposed by (Krishnanand & (Karaboga, 2005) is less complex when
Ghose, 2005). GSO mimics the behavior of compared with previous bee optimization
luminescent glowworms in nature. algorithms (Bolaji, Khader, Al-Betar, &
Awadallah, 2013) but converges poorly. (Sato
In this work a new algorithm called FBA that is & Hagiwara, 1997) reformulated the Genetic
inspired by the foraging behavior of bees is Algorithm (GA) to develop a new algorithm
proposed and implemented to improve the called the Bee System (BS). BS performs global
speed of convergence of bee-inspired search using GA operators and then improves
algorithms, avoid premature convergence as on local search by introducing new operators
well as balance exploitation with exploration. such as concentrated crossover and the pseudo-
simplex method.
2. LITERATURE REVIEW
I. Foraging Bee in Nature The mating behavior of bees is the inspiration
Honeybees are social insects or organisms that for the Mating Bee Optimization MBO
live together in well-organized colonies and can algorithm (Abbass, 2001). MBO algorithm
perform complex tasks in reasonable time with begins with one queen with no relatives, to a
ease. These tasks include controlling the colony of relatives with a single queen or
environment, division of labor, defense of nest multiple queens. MBO has been modified
and queen, nest construction, communications, several to form a new algorithms such as the
and foraging for food. The process of foraging Honey Bee Optimization (HBO) algorithm by
for food involves scouting, collection of pollen (Curkovic & Jerbic, 2007), Honey Bees Mating

100
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Optimization (HBMO) algorithm by (Haddad, that an intelligent forager forwarding strategy

Afshar, & Mariño, 2006) and the Fast Marriage significantly improves the quality of final
in Honey Bees Optimization (FMHBO) solutions and the convergence speed of ABC
algorithm by (Yang, Chen, & Tu, 2007). algorithms.

(Gao, Liu, & Huang, 2012) modified the ABC (Chen, Tianfield, & Du, 2021) proposed a novel
algorithm in order to improve its exploitation. bee-foraging learning PSO (BFL-PSO)
The new algorithm called ABC/Best searches algorithm that is inspired by the search
only around the fittest bee based on the last best mechanism of the artificial bee colony
solution. They employed a chaotic system and algorithm. The proposed BFL-PSO has three
opposition-based learning for improving the different search phases, namely: employed
speed of global convergence. learning, onlooker learning and scout learning.
The employed learning phase is the one-phase-
(Mathlouthi & Bouamama, 2016) integrated a based PSO search, while the onlooker learning
centralized and distributed technique called a phase exploits the region around promising
local optimum detector to an algorithm inspired solutions, and the scout learning phase
by marriage in honeybees. The local detector introduces new diversity by re-initializing
enhanced finding the local optimum. (Li & stagnant particles. The proposed BFL-PSO is
Yang, 2016) proposed a variant of ABC. They evaluated on the CEC2014 benchmark suite,
introduced a memory mechanism that aids and compared with state-of-the-art PSO and
artificial bees by memorizing their best foraging artificial bee colony algorithms The
experience so far. (Pan, 2016) hybridized ABC experimental results show BFL-PSO to be
and GA to develop a self-adaptive algorithm competitive in performance and the accuracy of
with a dual population of independently its solutions.
evolving bees that exchange information
through information entropy that ensures It is helpful to study and compare various
diversity and accelerates convergence. versions of bee inspired metaheuristics to
enable the selection of these algorithms in the
(Pan, 2016) proposed a hybrid, self-adaptive optimization tasks and the refinement and
genetic-bee colony algorithm based on development of new variants. (Solgi &
information entropy. The algorithm evolved Loáiciga, 2021) identifies seven basic or root
two populations of bees independently but algorithms applied to solve continuous
allowed the exchange of information between optimization problems, namely: Bee System
bees in the two populations using entropy to (BS), Mating Bee Optimization (MBO), Bee
maintain population diversity and accelerate the Colony Optimization (BCO), Bee Evolution
evolution process. Under analysis it was found for Genetic Algorithms (BEGA), Bee
that this strategy accelerated the emergence of Algorithm (BA), Artificial Bee Colony (ABC),
fitter individuals by competition between the and Bee Swarm Optimization. They ranked
populations performs better in complex these algorithms by performance and
function optimization problems. convergence efficiency and found ABC,
(Aslan, Karaboga, & Badem, 2020) modeled BEGA, and MBO to be the most efficient.
the complex behavior of foraging bees in detail They discussed the strengths and shortcomings
– how they pass through the dance area and how of each algorithm and explained the variations
long they performed their dance to attract observed in the convergence rate of these
onlooker bees – then adapted it to ABC to algorithms.
develop a new variant of ABC, termed the
intelligent forager forwarding ABC (iff-ABC). 3. THE FORAGING BEE
They analyzed the contribution of the intelligent OPTIMIZATION ALGORITHM (FBA)
forager forwarding strategy on the performance The Foraging Bee (Optimization) Algorithm
of ABC algorithms by evaluating its (FBA) is inspired by the foraging behaviour of
performance on the CEC benchmark suite and bees for pollen and nectar, and the collective
comparing it with the performance of different natural selection of more profitable nectar
variants of ABC. The results obtained showed sources over poor ones. The FBA algorithm is

101
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

developed. In FBA, the colony consists of bees, K number of scouts added as recruits
termed foragers, who scout for flowers that are during each waggle phase
rich sources of pollen and nectar in a patch of 𝑃𝑝𝑟𝑜𝑏 is the search space
the problem space in the work phase, then return 𝛽 is fraction of resource rich flowers
to the colony during the withdraw phase to for estimating the attractor of a patch
communicate their findings using dance in the p is the propensity of bees when
waggle phase. The FBA pseudocode is listed exploring patches
below as follows: 2 initialize M flowers in patches
set 𝑓𝑇 as the fitness of the fittest flower
Bee initialize bees randomly in patch 𝑃
A bee 𝑏𝑖 is modeled by the tuple 𝐵 = termcond ← false
𝐵(𝑥𝐵 , 𝑓𝐵 , 𝐷, 𝑃) where 𝑥𝐵 is the vector n ← 1, k ← 0
representing the current position of the bee, 3 while true
𝑓𝐵 ← 𝑓(𝑥𝐵 ) is the fitness of the current position // WORK PHASE
of the bee, 𝐷 is the direction of the bee and 𝑃 is for 𝑖 = 1 to T
the patch in which the bee is initialized. Each move bee 𝑏𝑖
bee makes a foraging move in time 𝑡 + 1 in if 𝑓(𝑏𝑖 ) < 𝑓𝑇 mark 𝑏𝑖 with flower
dimension 𝑗 as follows: 𝐹𝑀+𝑛 and increment n
if 𝑛 < 𝑁 then continue
𝑥𝑗 (𝑡 + 1) = 𝑥𝑗 (𝑡) // WITHDRAW PHASE
+ 𝑝𝑟1 (𝑑𝑗+ {𝑈𝑗 − 𝑥𝑗 (𝑡)} (1) termcond ← GET-TERMCOND( )
+ 𝑑𝑗− {𝑥𝑗 (𝑡) − 𝐿𝑗 }) if termcond then
return fittest flower as optimum
where 𝑟1 is a random number between 0 and 1, n←1
𝑈𝑗 𝑎𝑛𝑑 𝐿𝑗 are the upper and lower bounds in set 𝑓𝑇 as the fitness of the fittest
dimension 𝑗 of patch 𝑃, 𝑝 is the propensity of flower
the bee. The direction vector 𝐷 is a unit vector // WAGGLE PHASE
indicating the current direction of the foraging select best M flowers in 𝑃𝑝𝑟𝑜𝑏
bee. Bees make decisions before moving in Estimate promising patch using
direction 𝐷 by determining the direction 𝑑𝑗± to selected flowers 𝑃𝑏𝑒𝑠𝑡
Determine the location of attractors in
move in each dimension 𝑗 using the random
each patch.
variable 𝑟2 ~𝑈(0,1). Assume that the bee is
increment k
moving in direction 𝑑𝑗+ in time 𝑡. The decision initialize 𝑘 bees (recruits) randomly
to continue in direction 𝑑𝑗+ is determined by in 𝑃𝑏𝑒𝑠𝑡
initialize other bees 𝐵𝑝𝑜𝑝 − 𝑘 (scouts)
|𝑎𝑗 − 𝑥𝑗 | in 𝑃𝑝𝑟𝑜𝑏
𝑟2 < (2)
𝑈𝑗 − 𝐿𝑗
where 𝐴 = (𝑎1 , 𝑎2 , … , 𝑎𝑛 ) is the bee attractor in The propensity 𝑝 determines how bees explore
each patch. If (2) is true and 𝑐𝑗 is in direction of or exploit a patch. Lower values of 𝑝 favors
𝑈𝑗 , then 𝑑𝑗+ = 1 and 𝑑𝑗− = 0 otherwise 𝑑𝑗− = 1 exploitation over exploration. The continuous
and 𝑑𝑗+ = 0. reduction in the spatial dimensions of 𝑃𝑏𝑒𝑠𝑡
allows the exploitation of promising patches by
recruits while scouts continue to explore the
Algorithm 1: Foraging Bee Optimization
entire problem space. The bee search equation
Algorithm
guides the bees to forage only within the patch
1 set the following parameters
in which they are initialized making exploration
𝐵𝑝𝑜𝑝 is the population of bees and exploitation explicit processes guided by
M is the minimum population of patches 𝑃𝑝𝑟𝑜𝑏 and 𝑃𝑏𝑒𝑠𝑡 .
flowers
N is the minimum population of newly
discovered flowers

102
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Flower Start

A flower 𝐹 is modeled by the tuple 𝐹 =

𝐹(𝑥𝐹 , 𝑓𝐹 ), where 𝑥𝐹 is the vector representing Work
Phase
the current position of the flower 𝐹 and 𝑓𝐹 is its
fitness. The lower value of 𝑓𝐹 the richer the Withdraw
flower as a source nectar and pollen to bees. Phase

Is
Patch Termination
Stop
Patches are modeled by the tuple 𝑃 =
Yes
Condition?

𝑃(𝐿, 𝑈, 𝐴) where 𝐿 is the vector that represents

the lower limit of the patch in all dimensions, 𝑈 No

Waggle
is the vector that represents the upper limit of Phase

the patch in all dimensions and 𝐴 is the bee

attractor. 𝑃𝑝𝑟𝑜𝑏 is initialized with 𝑀 + 𝑁 or
more flowers while 𝑃𝑏𝑒𝑠𝑡 is estimated with 𝑀 Fig. 1. Flowchart of FBA
best flowers. The point attractor of bees in a
patch is the centroid the best flower 𝑓𝑇 and a 4. RESULT AND DISCUSSION
fraction 𝛽 ≅ 0.5 of the other flowers in the The FBA algorithm was run on standard
patch. Candidate flowers for bee attractor benchmark test function; these functions were
computation selected using the roulette operator presented in Table 1 as equations (3) to (17) and
[]. Unlike GA flowers are selected without their properties a tabulated in Table 2. They
replacement, i.e., a candidate flower cannot be were carefully chosen to test FBA’s capacity to
selected more than once. It is important that the solve problems with diverse properties and
point attractor of bees in patches 𝑃𝑝𝑟𝑜𝑏 and varying levels of difficulty. 𝑓1 to 𝑓3 are simple
unimodal functions while 𝑓4 to 𝑓15 are
𝑃𝑏𝑒𝑠𝑡 are not coincident at the early stages of
multimodal functions with local minima
search. Observe that all points in 𝑃𝑏𝑒𝑠𝑡 are
ranging from a few hundred to millions. The
interior points of 𝑃𝑝𝑟𝑜𝑏 .
performance of FBA on each test function is
benchmarked with 30 trials of 50000 function
Foraging Bee Algorithm evaluations.
The flowchart in Fig. 1 highlight phases FBA.
It begins with the initialization of search
10
parameters and objects such as bees, patches, 9
8
and flowers. This is followed by the work phase 7
where scout bees and recruits search for new 6
5
resource rich flowers. During the withdraw 4
3
phase critical parameters are reset and the GET- 2
1
TERMCOND method determines if an 0
approximate solution has been found or the -1
-2
maximum number of objective function -3
-4
evaluation has been exceeded. If the algorithm -5
-6
does not stop it enters the waggle phase where -7
-8
information is shared; 𝑃𝑏𝑒𝑠𝑡 is initialized or -9
recalibrated and scouts are recruited to exploit -10
-11
the patch. The algorithm repeats the work,
withdraw and waggle phases until it terminates BEST VALUE AVERAGE VALUE
in the withdrawal phase. WORST VALUE

Fig. 2. Graph of FBA results

The overall performance based on the best,

average, and worst-case error rates, standard
deviation, and success rate of each test function

103
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

is tabulated in Table 3 and shown graphically in determined by an error of at least four leading
Fig. 2. The success rate of each function is also zeros (E-04).
shown in Fig. 5. The success of any run is

Table 1. Benchmark test functions

1) Sphere function
𝑛

𝑓1 (𝑥) = ∑ 𝑥𝑖2 (3)

𝑖=1

2) Schwefel P2.22 function

𝑛 𝑛

𝑓2 (𝑥) = ∑|𝑥𝑖 | + ∏|𝑥𝑖 | (4)

𝑖=1 𝑖=1

3) Rosenbrock’s function
𝑛−1

𝑓3 (𝑥) = ∑{100(𝑥𝑖2 − 𝑥𝑖+1 )2 + (𝑥𝑖 − 1)2 } (5)

𝑖=1

4) Ackley F1
𝑛 𝑛
1 1
𝑓4 (𝑥) = −20 exp (−0.2√ ∑ 𝑥𝑖2 ) − exp ( ∑ cos(2𝜋𝑥𝑖 )) + 20 + 𝑒 (6)
𝑛 𝑛
𝑖=1 𝑖=1

5) Goldstein-Price
𝑓5 (𝑥) = {1 + (𝑥1 + 𝑥2 + 1)2 (19 − 14𝑥1 + 3𝑥12 − 14𝑥2 + 6𝑥1 𝑥2 + 3𝑥22 )}
(7)
× {30 + (2𝑥1 − 3𝑥2 )2 (18 − 32𝑥1 + 12𝑥12 + 48𝑥2 − 36𝑥1 𝑥2 + 27𝑥22 )}
6) Penalized Function P8
𝐷−1 𝐷
𝜋
𝑓6 (𝑥) = {10 sin2 (𝜋𝑦𝑖 ) + ∑{1 + 10 sin2 (𝜋𝑦𝑖+1 )} + (𝑦𝑑 − 1)2 } + ∑ 𝜇(𝑥𝑖 , 10,100,4) (8)
𝑥
𝑖=1 𝑖=1

7) Penalized Function P16

𝑓7 (𝑥) = 0.1 {sin2 (3𝜋𝑥𝑖 )

𝑛−1

+ ∑(𝑥𝑖 − 1)2 {1 + 10 sin2 (3𝜋𝑥𝑖+1 )} + (𝑥𝑑 − 1)2 {1 + 10 sin2 (2𝜋𝑥𝐷 )}} (9)
𝑖=1
𝐷

+ ∑ 𝜇(𝑥𝑖 , 5,100,4)
𝑖=1

8) Schaffer’s F6 function
sin2 (√∑𝑛𝑖=1 𝑥𝑖2 ) − 0.5
𝑓8 (𝑥) = 0.5 + (10)
{1 + 0.001(∑𝑛𝑖 𝑥𝑖2 )}2
9) Shekel 5 function
5
1
𝑓9 (𝑥) = − ∑ 2 (11)
4
𝑖=1 ∑𝑗=1(𝑥𝑖 − 𝑎𝑖𝑗 ) + 𝑐𝑖
10) Shekel 7 function
7
1
𝑓10 (𝑥) = − ∑ 2 (12)
𝑖=1 ∑4𝑗=1(𝑥𝑖 − 𝑎𝑖𝑗 ) + 𝑐𝑖
11) Shekel 10 function

104
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

10
1
𝑓11 (𝑥) = − ∑ 2 (13)
𝑖=1 ∑4𝑗=1(𝑥𝑖 − 𝑎𝑖𝑗 ) + 𝑐𝑖
4 4 4 4 0.1
1 1 1 1 0.2
8 8 8 8 0.2
6 6 6 6 0.4
0.4
𝐴 = [𝑎𝑖𝑗 ] = 3 7 3 7 𝐶 = [𝑐𝑖 ] =
2 9 2 9 0.6
5 5 3 3 0.3
8 1 8 1 0.7
6 2 6 2 0.5
[7 3.6 7 3.6] [0.5]
12) Six-Hump Camelback
𝑥14 2
𝑓12 (𝑥) = (4 − 2.1𝑥12 + ) 𝑥 + 𝑥1 𝑥2 + (−4 + 4𝑥22 )𝑥22 (14)
3 1
13) Schwefel P2.6 function
𝑛

418.9829𝑛 − ∑ 𝑥𝑖 sin (√|𝑥𝑖 |) (15)

𝑖=1

14) Griewank’s function

𝑛 𝑛
1 𝑥𝑖
𝑓14 (𝑥) = 1 + ∑ 𝑥𝑖2 − ∏ cos ( ) (16)
4000 √𝑖
𝑖=1 𝑖=1

15) Rastrigin’s function

𝑛

𝑓15 (𝑥) = ∑(𝑥𝑖2 − 10 cos(2𝜋𝑥𝑖 ) + 10) (17)

𝑖=1

Table 2. Properties of benchmark test functions

Name Feasible Bounds 𝑛 Optimum, 𝒙∗ 𝒇(𝒙∗ )
𝑓1 Sphere [−100, 100]𝑛 5 0𝑛 0
𝑓2 Schwefel P2.22 [−500, 500]𝑛 5 420.9687𝑛 0
𝑓3 Rosenbrock’s [−100, 100]𝑛 5 1𝑛 0
𝑓4 Ackley’s F1 [−32.768, 32.768] 5 0𝑛 0
𝑓5 Goldstein-Price [−2, 2] 2 (0, −1) 0
𝑓6 Penalized F8 [−50, 50] 5 −1𝑛 0
𝑓7 Penalized P16 [−50, 50] 5 1𝑛 0
𝑓8 Schaffer F6 [−100, 100]𝑛 2 0𝑛 0
𝑓9 Shekel 5 [0, 10]𝑛 4 4.0𝑛 -10.1499
𝑓10 Shekel 7 [0, 10]𝑛 4 4.0𝑛 -10.3999
𝑓11 Shekel 10 [0, 10]𝑛 4 4.0𝑛 -10.5319
𝑓12 Six-Hump Camel [−5, 5]𝑛 2 (-0.0898, 0.7126), -1.0316
(0.0898, -0.7126)
𝑓13 Schwefel P2.6 [−500, 500]𝑛 5 420.9687𝑛 0
𝑓14 Griewank [−600, 600]𝑛 5 0𝑛 0
𝑓15 Rastrigin [−5.12, 5.12] 5 0𝑛 0

105
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Table 3. The summary results obtained by the FBA algorithms for 30 runs
Func. Best Value Average Value Worst Value Std. Dev. Success Rate
𝑓1 1.0686E-119 1.7687E-16 5.2875E-15 9.6525E-16 100
𝑓2 4.6843E-33 3.7497E-04 7.0707E-03 1.4766E-03 93.33
𝑓3 2.8994E+00 1.8798E+04 2.2862E+04 4.4271E+03 0
𝑓4 0.0000E+00 7.1304E-08 2.1232E-06 3.8756E-07 100
𝑓5 -9.5923E-14 5.4712E-05 1.6414E-03 2.9967E-04 96.67
𝑓6 2.1903E-11 3.8512E-02 9.2391E-01 1.6801E-01 43.33
𝑓7 1.4096E-14 6.2899E-05 1.5010E-03 2.7692E-04 96.67
𝑓8 3.3695E-13 3.7001E-04 2.4989E-03 5.6688E-04 90
𝑓9 -3.8281E-06 2.7995E+00 7.6638E+00 2.8028E+00 33.33
𝑓10 -1.2173E-04 1.0171E+00 6.4562E+00 1.8296E+00 60
𝑓11 -1.2609E-04 1.2683E+00 7.7326E+00 2.2092E+00 66.67
𝑓12 -3.0562E-08 9.4618E-03 1.0871E-01 2.0714E-02 33.33
𝑓13 3.5809E+01 1.4418E+02 2.0573E+02 4.3292E+01 0
𝑓14 2.1281E-02 7.9438E-02 1.2790E-01 2.6691E-02 0
𝑓15 1.7127E+00 3.2470E+00 6.2162E+00 1.1694E+00 0

Fig. 3(a) to (k) show successful runs of FBA on and 𝑓15) benchmark functions out of the fifteen
11 benchmark test functions for which it tested on. Three (𝑓6, 𝑓9 , and 𝑓12) were below
converges, and successfully returns at least once fifty percent while the remaining eight ranges
an approximate solution to the optimum with from sixty to hundred percent.
error rates less the 1E-08. Fig. 4 (a) to (d) on the
other hand are unsuccessful runs of FBA on 4 Results in Table 5 show that the comparison
benchmark test functions. between FBA and PSO on all 10 benchmark test
functions is statistically significant. Results also
FBA is compared ABC and PSO using T-test. show that 7 out of the 10 benchmark tests
Table 4 shows the mean and standard error of between FBA and ABC are statistically
FBA, ABC and PSO on the test function while significant. In nine of the ten statistically
Table 5 tabulates the results of the T-test and significant benchmark test FBA performed
indicates test functions in which the better than PSO while in two of the seven
performance of FBA is statistically significant statistically significant test FBA performed
when compared with both ABC and PSO. FBA better than ABC.
did not return any success for four (𝑓3, 𝑓13, 𝑓14,

1
0,9 1
0,8 0,9
0,7 0,8
0,7
FITNESS

0,6
FITNESS

0,6
0,5
0,5
0,4
0,4
0,3
0,3
0,2
0,2
0,1
0,1
0 0
0 250 500 750 1000 0 1000 2000 3000 4000 5000
FUNCTION EVALUATION FUNCTION EVALUATION

Fig. 3(a). Sphere Fig. 3(b). Schwefel P2.22

106
IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

1 13
0,9 12
11
0,8 10
0,7 9

FITNESS
FITNESS

0,6 8
0,5 7
6
0,4 5
0,3 4
0,2 3
0,1 2
1
0 0
0 250 500 750 1000 1250 1500 0 100 200 300 400 500 600 700 800
FUNCTION EVALUATION FUNCTION EVALUATION

Fig. 3(c). Ackley Fig. 3(d). Goldstein-price

1
0,9
0,8
0,7
FITNESS

0,6
0,5
0,4
0,3
0,2
0,1
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
FUNCTION EVALUATION

Fig. 3(e). Penalized function P8

1 0,5
0,9
0,8 0,4
0,7
FITNESS

0,6
FITNESS

0,3
0,5
0,4 0,2
0,3
0,2
0,1
0,1
0
0
0 250 500 750 1000 1250 1500 1750 2000
0 250 500 750 1000 1250 1500 1750 2000
FUNCTION EVALUATION FUNCTION EVALUATION

Fig. 3(f). Penalized function P16 Fig. 3(g). Schaffer F6

-9 -10
0 250 500 750 1000 1250 1500 1750 2000-9,1 0 250 500 750 1000 1250 1500 1750 2000
-9,2
-9,3 -10,1
-9,4
-9,5 -10,2
FITNESS
FITNESS

-9,6
-9,7
-9,8 -10,3
-9,9
-10 -10,4
-10,1
-10,2
-10,5
-10,3 FUNCTION EVALUATION
FUNCTION EVALUATION
Fig. 3(h). Shekel 5 Fig. 3(i). Shekel 7

107
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

-10 FUNCTION EVALUATION

0 250 500 750 1000 1250 1500 1750 2000
-10,1 0 50 100 150 200 250 300 350 400 450 500
0
-10,2 -0,1
-0,2

FITNESS
-0,3
-10,3 -0,4

FITNESS
-0,5
-10,4 -0,6
-0,7
-10,5 -0,8
-0,9
-1
-10,6
FUNCTION EVALUATION -1,1
Fig. 3(j). Shekel 10 Fig. 3(k). Six-Hump Camel

30
25
20
FITNESS

15
10
5
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
FUNCTION EVALUATION

Fig. 4(a). Rosenbrock

1
0,9
0,8
0,7
FITNESS

0,6
0,5
0,4
0,3
0,2
0,1
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
FUNCTION EVALUATION
Fig. 4(b). Griewank

100
0 90
80
-50 0 10 20 30 40 50
70
FITNESS

60
FITNESS

-100 50
-150 40
30
-200 20
10
-250 0
FUNCTION EVALUATION 0 500 1000 1500 2000 2500 3000
FUNCTION EVALUATION

Fig. 4(c). Schwefel P2.6 Fig. 4(d). Rastrigin

108
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Fig. 5. Histogram of success rates

Table 4: The Comparison between FBA, PSO, and ABC

FBA ABC PSO
Function
Mean Std. Error Mean Std. Error Mean Std. Error
Sphere 1.77E-16 ±1.76E-16 6.99E-10 ±1.08E-10 2.75E+00 ±4.48E-02
Schwefel P2.22 3.75E-04 ±2.70E-04 2.36E-06 ±1.52E-07 5.45E+00 ±1.43E-01
Rosenbrock 1.87E-03 ±8.08E+02 3.93E-02 ±5.68E-03 5.46E+01 ±2.85E+00
Ackley F1 7.13E-08 ±7.08E-08 1.02E-05 ±4.45E-03 2.02E+01 ±8.20E-03
Penalized F. P8 3.85E-02 ±3.07E-02 1.60E-11 ±3.56E-12 8.86E+00 ±3.83E-01
Penalized P16 6.29E-05 ±5.06E-05 3.72E-09 ±3.27E-10 1.22E-01 ±2.10E-03
Schaffer F6 3.70E-04 ±1.03E-04 2.07E-03 ±7.14E-04 - -
Shekel 7 -9.38E+00 ±3.34E-01 -1.04E-01 ±2.88E-16 5.31E+00 ±6.05E-06
Shekel 10 -9.26E+00 ±4.03E-01 -1.05E-01 ±6.19E-15 5.40E+00 ±6.19E-06
Griewank 7.94E-02 ±4.87E-03 8.73E-09 ±2.68E-09 1.01E+00 ±3.10E-03
Six-Hump Camel 9.46E-03 ±2.07E-02 - - 1.77E+00 ±2.89E-01

Table 5: The T-test between FBA/PSO and FBA/ABC

T-test FBA/PSO T-test FBA/ABC
Critical Critical
Function Value Significant Value Significant
Value Value
Sphere 61.5223 <0.00001 YES 6.4781 <0.00001 YES
Schwefel P2.22 38.2318 <0.00001 YES 1.3829 0.0885 YES
Rosenbrock 2.2662 0.0154 YES 2.3258 0.0135 YES
Ackley F1 2472.81 <0.00001 YES 0.0023 0.4991 NO
Penalized F. P8 25.0648 <0.00001 YES 1.2555 0.1095 NO
Penalized P16 59.4977 <0.00001 YES 1.2440 0.1116 NO
Schaffer F6 - - - 2.7852 0.0046 YES
Shekel 7 -44.8228 <0.00001 YES 3.0444 0.0024 YES
Shekel 10 -36.3861 <0.00001 YES 3.1452 0.0019 YES
Griewank -525.4461 <0.00001 YES 16.3014 <0.00001 YES
Six-Hump Camel 6.5879 <0.00001 YES - - -

109
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

V. CONCLUSION REFERENCES
The Foraging Bee Optimization (FBA) Abbass, H. A. (2001). MBO: Marriage in
algorithm is a swarm intelligence optimization Honey Bees Optimization-A
algorithm, which has a new unique approach Haplometrosis Polygynous Swarming
inspired by the characteristics and intelligent
Approach. Proceedings of the 2001
behavior displayed by the swarm of foraging
bees for solving optimization problems. The IEEE Congress on Evolutionary
algorithm mimics bee colonies by organizing Computation (IEEE Cat. No.
search into three phases: work – when bees 01TH8546) Vol. 1., 207-214.
forage in patches and discover and exploit new Alizadehsani, R., Roshanzamir, M., Izadi, N.
resource-rich flowers; withdraw – when bees H., Gravina, R., Kabir, H. D.,
return to the colony and reset for the next work Nahavandi, D., . . . Fortino, G. (2023).
phase; and waggle – when bees share
information about locations containing resource Swarm Intelligence in Internet of
rich flowers. Medical Things: A Review. Sensors,
23(3), 1466.
The algorithm increased the speed of doi:https://doi.org/10.3390/s2
convergence and balance exploration and Altshuler, Y. (2023). Recent Developments in
exploitation by positioning flowers at the the Theory and Applicability of
extreme of a rectangular workspace that must be
Swarm Search. Entropy, 25(5).
scooped by the bees with a propensity that
ensures thorough exploration. Exploitation is Aslan, S., Karaboga, D., & Badem, H. (2020).
achieved by a spatial reduction of the best patch A New Artificial Bee Colony
subspace over several 3W cycles. Unlike PSO, Algorithm employing Intelligent
FBA avoids stagnation and minimizes the Forager Forwarding Strategies.
possibility of premature convergence that Applied Soft Computing, 96.
occurs when algorithms are guided by Belgrade, U. (2015, September 16). Bee
exemplars, honeybees in FBAs are guided by
Colony Optimization. Retrieved
attractors which shift as new flowers are
discovered. The proposed algorithm was tested September 16, 2015, from
on fifteen standards benchmarks of which three http://www.sf.bg.ac.rs/index.php/en-
are unimodal while the remaining are complex GB/research-fields/814-bee-colony-
multimodal spaces with millions of local optimization-bco
optima. Bolaji, A. L., Khader, A. T., Al-Betar, M. A.,
& Awadallah, M. A. (2013). Artificial
The algorithm was compared with two state-of-
the-art algorithms PSO and ABC, and the Bee Colony Algorithm, Its Variants
statistically significant result shows that FBA is and Applications. A Survey, Journal of
more efficient than PSO while being Theoretical and Applied Information
competitive with ABC. FBA has been tested Technology, 47(2), 434-459.
extensively in this work, but more experiments Chen, X., Tianfield, H., & Du, W. (2021). Bee-
need to be done to tune the parameters of FBA foraging Learning Particle Swarm
for solving more complex test functions at
Optimization. Applied Soft Computing,
higher dimensions. In addition, adaptive
variants of FBA should be developed to reduce 102. doi:10.1016/j.asoc.2021.107134
the number of parameters that require tuning for Cruz, D. P., Maia, R. D., & de Castro, L. N.
high performance. Finally, an information- (2021). A Framework for the Analysis
sharing mechanism will be developed to reduce and Synthesis of Swarm Intelligence
the overall complexity of the search algorithm. Algorithms. Journal of Experimental
& Theoretical Artificial Intelligence,
33, 659-681.

110
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Curkovic, P., & Jerbic, B. (2007). Honey-bees Erciyes University, engineering

optimization algorithm applied to path faculty, computer engineering
planning problem. International department., Kayseri, Turkiye.
journal of simulation modelling, 6(3), Kaswan, K. S., Dhatterwal, J. S., & Kumar, A.
154-165. (2023). Swarm Intelligence: An
Dorigo, M., Colorni, A., & Maniezzo, V. Approach from Natural to Artificial.
(1991). Positive feedback as a search John Wiley & Sons.
strategy. Technical Report 91-016, Kennedy J. and Eberhart, R. (1995). Particle
Politecnico di Milano, Dipartimento di swarm optimization. Proceedings of
Elettronica, Milan, Italy. the IEEE International Conference on
Eberhart, R. C., Shi, Y., & Kennedy, J. (2001). Neural Networks, 4, pp. 1942–1948.
Swarm intelligence. Elsevier. Krishnanand , K. N., & Ghose, D. (2005).
Engelbrecht, A. P. (2007). Computational Detection of multiple source locations
intelligence: an introduction (2nd ed.). using a glowworm metaphor with
Pretoria, South Africa: John Wiley & applications to collective robotics. in
Sons. Proceedings of the IEEE Swarm
Fakhermand, S. M., & Derakhshani, A. (2023). Intelligence Symposium (SIS ’05), (pp.
Design Optimization of Soil-Metal 84–94). Pasadena, California.
Composite Arch Bridges: Recent Kumar, A., Chatterjee, J. M., Payal, M., &
Swarm Intelligence Applications. Rathore, P. S. (2022). Revolutionizing
Iranian Journal of Science and the Internet of Things with Swarm
Technology, Transactions of Civil Intelligence. System Assurances, 403-
Engineering, 47(1), 373-387. 436. doi:10.1016/B978-0-323-90240-
Gao, W., Liu, S., & Huang, L. (2012). A 3.00023-0
global best artificial bee colony Li, X., & Yang, G. (2016). Artificial bee
algorithm for global optimization. colony algorithm with memory.
Journal of Computational and Applied Applied Soft Computing, 41, 362-372.
Mathematics, 236(11), 2741-2753. Mathlouthi , I., & Bouamama, S. (2016). A
Haddad, O. B., Afshar, A., & Mariño, M. A. family of honey-bee optimization
(2006). Honey-Bees Mating algorithms for Max-CSPs.
Optimization (HBMO) Algorithm. A International Journal of Knowledge-
New Heuristic Approach for Water based and Intelligent Engineering
Resources Optimization. Water Systems, 19(4), 215-224.
Resources Management, 20(5), 661- Pan, X. (2016). Genetic-bee Colony Dual-
680. population Self-adaptive Hybrid
Holland, J. H. (1975). Adaption in Natural and Algorithm Based on Information
Artificial System. . MIT Press. Entropy. Scientific Bulletin of National
Janaki, M., & Geethalakshmi, S. N. (2022). A Mining University, 1(1), 116.
Review of Swarm Intelligence-Based Pham, D. T., Ghanbarzadeh, A., Koç, E., Otri,
Feature Selection Methods and Its S., Rahim, S., & Zaidi, M. (2005). The
Application. Soft Computing for Bees Algorithm – A Novel Tool for
Security Applications: Proceedings of Complex Optimization Problem. In D.
ICSCS 2022, 435-447. T. Pham, E. E. Eldukhri, & A. J.
Karaboga, D. (2005). An Idea Based on Honey Soroka (Ed.), Intelligent Production
Bee Swarm for Numerical Machines and Systems (p. 454).
Optimization. Technical report-tr06, Elsevier Science Ltd.

111
 IJIEM (Indonesian Journal of Industrial Engineering & Management) Vol 4 No 2 June 2023, 99-112

Sato, T., & Hagiwara, M. (1997). Bee System: continuous spaces. Journal of Global
Finding Solution by a Concentrated Optimization, 11(4), 341–359.
Search. IEEE International Teodorovic, D., & Dell’orco, M. (2005). Bee
Conference on Computational Colony Optimization—A Cooperative
Cybernetics and Simulation (pp. 3954- Learning Approach to Complex
395). Orlando, FL, USA: IEEE. Transportation Problems. Proceedings
Schumann, A. (. (2020). Swarm Intelligence: of the 16th Mini-EURO Conference on
From Social Bacteria to Humans. Advanced OR and AI Methods in
CRC Press. Transportation, (pp. 51-60). Poznan.
Selvaraj, S., & Choi, E. (2020). Survey of Retrieved September 13-16, 2005
swarm intelligence algorithms. 3rd Tzanetos, A., & Dounias, G. (2020). A
International Conference on Software Comprehensive Survey on the
Engineering and Information Applications of Swarm Intelligence
Management, (pp. 69-73). and Bio-Inspired Evolutionary
Shahzad, M. M., Saeed, Z., Akhtar, A., Strategies. In G. Tsihrintzis, & L. Jain,
Munawar, H., Yousaf, M. H., Baloach, Machine Learning Paradigms.
N. K., & F, H. (2023). A Review of Learning and Analytics in Intelligent
Swarm Robotics in a NutShell. Systems (Vol. 18, pp. 337-378). Cham:
Drones, 7(4), 269. Springer.
Solgi, R., & Loáiciga, H. A. (2021). Bee- doi:https://doi.org/10.1007/978-3-030-
Inspired Metaheuristics for Global 49724-8_15
Optimization: A Performance Yang, C., Chen, J., & Tu, X. (2007).
Comparison. Artificial Intelligence Algorithm of Fast Marriage in Honey
Review, 54(7), 4967-4996. Bees Optimization and Convergence
doi:10.1007/s10462-021-10015-1 Analysis. In Proceedings of the IEEE
Storn, R., & Price, K. V. (1997). Differential International Conference on
evolution—a simple and efficient Automation and Logistics (pp. 1794–
heuristic for global optimization over 1799). Jinan, China: ICAL.

112