TWO DISCRETE BINARY VERSIONS OF AFRICAN BUFFALO OPTIMIZATION METAHEURISTIC

Dhinaharan Nagamalai et al. (Eds) : ACSIT, ICITE, SIPM - 2018
pp. 33–46, 2018. © CS & IT-CSCP 2018 DOI : 10.5121/csit.2018.80803
TWO DISCRETE BINARY VERSIONS OF
AFRICAN BUFFALO OPTIMIZATION
METAHEURISTIC
Amira GHERBOUDJ
MISC laboratory, NTIC faculty, AbdelhamidMehri University Constantine 2,
Algeria
ABSTRACT
African Buffalo Optimization (ABO) is one of the most recent swarms intelligence based
metaheuristics. ABO algorithm is inspired by the buffalo’s behavior and lifestyle. Unfortunately,
the standard ABO algorithm is proposed only for continuous optimization problems. In this
paper, the authors propose two discrete binary ABO algorithms to deal with binary optimization
problems. In the first version (called SBABO) they use the sigmoid function and probability
model to generate binary solutions. In the second version (called LBABO) they use some logical
operator to operate the binary solutions. Computational results on two knapsack problems (KP
and MKP) instances show the effectiveness of the proposed algorithm and their ability to
achieve good and promising solutions.
KEYWORDS
Optimization, metaheuristic, swarm intelligence, binary problems, African Buffalo
Optimization, knapsack problems
1. INTRODUCTION
Solving optimization problem is finding a solution of sufficient quality (i.e. optimal or near
optimal) from a set of solutions taking into account the constraints imposed and objectives to
meet. It is to maximize or minimize one or a set of fitness functions respecting constraints of the
treated problem. This research line that sought the attention of several research teams is
intrinsically linked to operational research and it uses mathematical and computer tricks.
Various Methods have been proposed to solve optimization problems. They are often classified
into two classes: exact methods and approximate methods. The prohibitive cost associated with
exact methods has excited researchers to use approximate methods. The investigation in the area
of approximate methods gave rise to a subclass of approximate methods called "Metaheuristics".
They are general and applicable methods on a wide range of optimization problems.
Metaheuristics based on swarm intelligence have built a very active trend over the last decade.
They are generally inspired by the collective behavior of some species in evolving in groups and

34 Computer Science & Information Technology (CS & IT)
solving their problems. These species gather in swarm to build a collective force that allows them
to surpass their very limited individual capacities.
The African Buffalo Optimization (ABO) is one of the most recent swarm intelligence based
Metaheuristics. It was proposed in 2015, by Julius BeneoluchiOdili et al [1]. ABO is inspired
from the behavior of African buffaloes in the vast African forests and savannahs [1]. The recent
applications of ABO Metaheuristic for optimization problems have shown its promising
effectiveness as it has been proven in [2], [3] and [4].The original ABO algorithm is a continuous
version, which solves only continuous problems. The aim of this paper is to propose two binary
versions of ABO algorithm to cope with binary optimization problems. The main difference
between the original version of ABO algorithm and the proposed binary versions is that, in the
original ABO algorithm, the solution is composed of a set of real numbers. While in the proposed
binary versions, the solution is composed of a set of bits.
The remainder of this paper is organized as follows. Section 2 presents an overview of African
Buffalo Optimization (ABO) Algorithm. The proposed algorithms (SBABO and LBABO) are
presented in section 3. Experimental results are presented and discussed in section 4, and a
conclusion and perspectives are provided in the fifth section of this paper.
2. AFRICAN BUFFALO OPTIMIZATION
In order to solve complex problems, ideas gleaned from natural mechanisms have been exploited
to develop heuristics. Nature inspired optimization algorithms has been extensively investigated
during the last decade paving the way for new computing paradigms such as neural networks,
evolutionary computing, swarm optimization, etc. The ultimate goal is to develop systems that
have the ability to learn incrementally, to be adaptable to their environment and to be tolerant to
noise. One of the recent developed bioinspired algorithms is the African Buffalo Optimization
algorithm.
The African Buffalo Optimization is inspired from the cooperative and competitive behavior of
buffaloes. ABO models the three characteristic behaviors of the African buffaloes that enable
their search for pastures. First is their extensive memory capacity. This enables the buffaloes to
keep track of their routes. The second attribute of the buffaloes is their cooperative
communicative ability whether in good or bad times. The third attribute of the buffaloes is their
democratic nature borne out of extreme intelligence. In cases where there are opposing calls by
members of the herd, the buffaloes have a way of doing an ‘election’ where the majority decision
determines the next line of action [1]. Furthermore, ABO algorithm models the two sounds for
communication that buffaloes use to exploit and explore the search space:
• The warning sound “waaa” with which they ask the herd to keep moving because the
present location is unfavorable, lacks pasture or is dangerous. This sound encourages the
buffaloes to explore the research space.
• The alert sound “maaa” with which they stay on the present location because it is holds
promise of good grazing pastures and is safe. This sound encourages the buffaloes to
exploit the research space
Algorithm1 presents a pseudo algorithm of the ABO method.

Computer Science & Information Technology (CS & IT) 35
Algorithm 1: ABO Algorithm
1. Objective function f(x) = (x1,x2,…xn)T
;
2. Initialization: randomly place buffaloes to nodes at the solution space;
3. Update the buffaloes fitness values using (1);
4. Update the location of buffalo k (bpmax.k and bgmax) using (2);
5. Is bgmax updating. Yes, go to 6. No, go to 2;
6. If the stopping criteria is not met, go back to algorithm step 3, else go to step 7;
7. Output the best solution.
The generation of new solutions is done by using equations 1 and 2.
).max.().max(.. 211 kkkkk wbplpwbglpmm −+−+=+
(1)
5.0
..
. 1
+−
+
=+
kk
k
mw
w (2)
Where:
• w.k and m.k presents the exploration and exploitation moves respectively of the kth
buffalo
(k=1,2,………..N) ;
• lp1 and lp2 are learning factors;
• bgmax is the herd’s best fitness;
• bpmax.k the individual buffalo’s best.
3. THE PROPOSED DISCRETE BINARY VERSIONS
Optimization problems can be classed into two main classes: continuous optimization problems
and discrete optimization problems. In continuous optimization problems, the solution is
presented by a set of real numbers. However, in discrete optimization problems, the solution is
presented by a set of integer numbers. Discrete binary optimization problems are a sub-class of
the discrete optimization problems class, in which a solution is presented by a set of bits. Many
optimization problems can be modelled as a discrete binary search space such as, flowshop
scheduling problem [5], job-shop scheduling problem [6], routing problems [7], KP [8] and its
variants such as the MKP [9], the quadratic KP [10], the quadratic multiple KP [11] and so one.
The original ABO algorithm operates in continuous search space. It gives a set of real numbers as
a solution of the handled problem. However, a binary optimization problem needs a binary
solution and the real solutions are not acceptable because they are considered as illegal solutions.
In the aim to extend the ABO algorithm to discrete binary areas, we propose in this paper tow
binary versions of ABO that we called SBABO and LBABO. The main objective of the SBABO
and LBABO algorithms is to deal with the binary optimization problems.

3.1 SBABO ALGORITHM
In the SBABO algorithm, we introduce a binarization phase of solutions in the core of the
original ABO in order to obtain a binary solution for the treated problem. The objective of this
phase (i.ebinarization) is to transform a solution xi from real area to binary area. To meet this
need, we propose to constrain the solution xi in the interval [0, 1] using the Sigmoid Function as
follows:
)1(
1
)( ixi
e
xS −
+
= (3)
Where S(xi) is the flipping chance of bit x′i. It presents the probability of bit x′i takes the value 1.
To obtain the binary solution x′i, we have to generate a random number from the interval [0,1] for
each dimension i of the solution x and compare it with the flipping chance S(xi) as mentioned
below in equation (4). If the random number is lower than the flipping chance of bit x′i, x′i takes
the value 1. Otherwise, x′i takes the value 0.
Consequently, having a solution xi encoded as a set of real numbers, the sigmoid function is used
to transform the solution xi into a set of probabilities that presents the chance for bit i to be
flipping. The flipping chance is then used to compute the binary solution x'i. Algorithm 2 presents
the SBABO algorithm an Algorithm 3 presents the binarization algorithm
Algorithm 2: SBABO Algorithm
2. Initialization: randomly place buffaloes to nodes at the solution space;
3.Update the buffalo’s fitness values using (1);
4.Get the binary buffaloes using the Binarization Algorithm;
5. Update the location of buffalo k (bpmax.k and bgmax) using (2);
7. If the stopping criteria is not met, go back to algorithm step 3, else go to step 8.
Algorithm 3:Binarization Algorithm
Input: Real solution presentation xi
Output: Binary solution presentation xi′
For (i = 1 to (problem size)) {
Calculate S(xi) using (3);
If (random number r <S(xi))
x′i=1;
Otherwise
x′i =0;
}

3.2 LBABO ALGORITHM
In the LBABO algorithm, we propose to start the search by a binary population (the solutions are
binary from the beginning) and replace the arithmetic operators used in the solution update
equations (i.e. 1 and 2) by logical operators as follow:
Coefficient can be lp1 or lp2. Algorithm 4 presents the LBABO algorithm
Algorithm 4. LBABO Algorithm
;
2. Initialization: randomly place buffaloes to nodes at the solution space using binary values;
3. Update the buffalo’s fitness values using (1) and logical operators;
4. Update the location of buffalo k (bpmax.k and bgmax) using (2) and logical operators;
6. If the stopping criteria is not met, go back to algorithm step 3, else go to step 7.
4. EXPERIMENTAL RESULTS
In order to investigate the performance of the proposed algorithms to solve hard binary
optimization problems, we used some knapsack problem benchmarks of two knapsack problem
versions: Single knapsack problem (KP) and multidimensional knapsack problem (MKP).
4.1 KP AND MKP PROBLEMS
The KP is a NP-hard problem [12]. Numerous practical applications of the KP can be found in
many areas involving resource distribution, investment decision making, budget controlling,
project selection and so one. The KP can be defined as follows: Assuming that we have a

knapsack with maximum capacity C and a set of N objects. Each object i has a profit pi and a
weight wi. The problem consists to select a subset of objects that maximize the knapsack profit
without exceeding the maximum capacity of the knapsack. The problem can be formulated as
[12]:
Many variants of the KP were proposed in the literature including the MKP. MKP is an important
issue in the class of KP. It is a NP-hard problem [13]. In the MKP, each item xi has a profit pi like
in the simple KP. However, instead of having a single knapsack to fill, we have a number M of
knapsack of capacity Cj (j = 1, …, M). Each xi has a weight wij that depends of the knapsack j
(example: an object can have a weight 3 in knapsack 1, 5 in knapsack 2, etc.). A selected object
must be in all knapsacks. The objective in MKP is to find a subset of objects that maximize the
total profit without exceeding the capacity of all dimensions of the knapsack. MKP can be stated
as follows [14]:
The MKP can be used to formulate many industrial problems such as the capital budgeting
problem, allocating processors and databases in a distributed computer system, cutting stock,
project selection and cargo loading problems [15]. Clearly, there are 2N
potential solutions for
these problems. It is obviously that KP and its variants are combinatorial optimization problems.
Several techniques have been proposed to deal with KPs [12]. However, it appears to be
impossible to obtain exact solutions in polynomial time. The main reason is that the required
computation grows exponentially with the size of the problem. Therefore, it is often desirable to
find near optimal solutions to these problems.
4.2 EXPERIMENTAL DATA
The proposed SBABO and LBABO algorithms were implemented in MATLAB R2014a. Using a
laptop computer running Windows 7, Intel(R) Core(TM) i3-3110M CPU@ 2.40 GHz, 2.40GHz,
4GB RAM. The used parameters in the experiments are:
• Population size: 40.
• Iterations: 300.
• lp1=0.7.
• lp2=0.5.

Several experiments were performed to assess the efficiency and performance of our algorithms.
In the first experiment, we have tested and compared our algorithms with a harmony search
algorithm (NGHS) on some small KP instances taken from [16]. In the second experiment, we
have used some big KP instances used in [8] to test and compare the proposed SBABO algorithm
with the Binary Particle Swarm Optimization algorithm (BPSO) [17] which has a common point
with the proposed SBABO algorithm. In fact, the two algorithms (SBABO and BPSO) used the
Sigmoid Function to generate the binary solution. The used instances are six different instances
with different problem sizes, in which the weights and profits are selected randomly. The
different problem sizes N are 120, 200, 500, 700, 900 and 1000. In these instances, the knapsack
capacity is calculated by using equation 9 [8]. The factor 3/4 indicates that about 75% of items
are in the optimal solution.
In the third experiment, we have evaluated the performance of our algorithms on some MKP
benchmarks taken from OR-Library. We have tested the proposed algorithms on some MKP
instances taken from seven benchmarks named mknap1. The obtained results are compared with
the exact solution (best known).
Finally, statistical tests of Freidman are carried out to test the significance of the difference in the
accuracy of each method in this experiment. And the performances of the proposed algorithms
(SBABO and LBABO) are also compared in terms of execution CPU time with the two problems
(KP and MKP).
4.3. RESULTS AND DISCUSSION
Table 1 shows the experimental results of our algorithms (SBABO and LBABO) and the
harmony search algorithm (NGHS) on ten KP tests with different sizes. The first column,
indicates the instance name, the second column indicates the problem size, i.e. number of objects.
The third and fourth columns indicate the obtained results by the SBABO and LBABO
algorithms and the last column indicates the obtained results by the NGHS algorithm.
Observation of the presented results in Table 1 indicates that the proposed discrete binary
algorithms (i.e: SBABO and LBABO) perform well than NGHS algorithm in F6 test. The
SBABO perform well than LBABO and NGHS algorithms in F8 test. And the three algorithms
have the same results in the other instances.
Table 1.Experimental results with small kp instances
Test Size SBABO LBABO NGHS
F1 10 295 295 295
F2 20 1024 1024 1024
F3 4 35 35 35
F4 4 23 23 23
F5 15 481.0694 481.0694 481.0694
F6 10 52 52 50
F7 7 107 107 107
F8 23 9767 9767 9767
F9 5 130 130 130
F10 20 1025 1025 1025

Table 2 shows the experimental results of SBABO, LBABO and BPSO algorithms on big KP
instances. The first column presents the problem size (i.e., instance). The second, third and fourth
columns present the obtained results by the BPSO, SBABO and LBABO algorithms respectively.
With each instance, the first line presents the best solutions and the second one presents the
solution averages. The presented results show that the SBABO and LBABO algorithms
outperform the BPSO algorithm, and the LBABO algorithm outperforms the SBABO algorithm.
The statistical Friedman test in Figure 1 presents a comparison of the BPSO, SBABO and
LBABO results. The LBABO algorithm ranks first in the Friedman test. The SBABO ranks
second and BPSO ranks third. This statistical test shows that there is a significant difference
between LBABO and BPSO algorithms.
Table 2. Experimental results with big kp instances.
Instance BPSO SBABO LBABO
120
200
500
4296
3840.8
7456
5703
13116
12471.2
4316
4088.09
6778
6480.56
14730
14396.11
4504
4357
7530
7284.22
16853
16174.25
700
18276 20501 23278
17097.4 19348.07 22530.4
900
22857 24767 30196
21736.6 24270.83 28864.5
1000
24933 27306 32948
24050 26607.3 31936.86
Whereas, the difference between LBABO and SBABO results is not statistically significant.
Consequently, the obtained results confirm that the proposed algorithms outperform the BPSO
algorithm and prove that the proposed algorithms give good results.
Table 3 shows experimental results of the proposed algorithms over 7 instances of MKP problem.
The first column indicates the instance index. The second and third column indicates the number
of object and knapsack dimension, respectively. The fourth, fifth and sixth columns indicate the
best known, the SBABO and the LBABO solutions, respectively. As we can see, the SBABO
algorithm is able to find the best solution of the six first instances from the 7 instances. The
LBABO algorithm is able to find the best solution of the five first instances from the 7 instances.
The SBABO algorithm outperforms the LBABO algorithm on the two last instances (6 and 7).
The statistical Friedman test in Figure 2 shows a comparison of the best known, SBABO and
LBABO results. The SBABO ranks second after best known and LBABO ranks third. This
statistical test shows that the difference between best known, LBABO and SBABO results is not
statistically significant. Consequently, the obtained results confirm and prove that the proposed
algorithms give good and promising results that can be considerably increased by the introduction
of some specified knapsack heuristic operators using problem specific knowledge.

Figure 1. Friedman test compares SBABO, LBABO and BPSO on big KP instances.
Table 3.Experimental results of MKP with mknap1 instances
N° N M Best Known SBABO LBABO
1 6 10 3800 3800 3800
2 10 10 8706,1 8706,1 8706,1
3 15 10 4015 4015 4015
4 20 10 6120 6120 6120
5 28 10 12400 12400 12400
6 39 5 10618 10618 10554
7 50 5 16537 16442 16371
Figure 2. Friedman test compares SBABO, LBABO and best known on MKP instances.

Table 4 and 5 show a comparison of average computation time with KP and MKP instances,
estimated by seconds, using a population of 40 solutions, 300 iterations with 5 executions of the
programs. The obtained results are schematized in Figures 3 and 4. In terms of computing time,
the obtained results do not show a big difference in execution time. In fact, in some instances
SBABO is faster and in others LBABO is faster. In general, the two algorithms converge in the
same interval time. This comes back to the fact that the two algorithms have the same body, only
the phase of binarization of the solution that differs.
Table 4.Comparative CPU time with KP instances.
Test Size SBABO LBABO
F1 10 2.50 2.75
F2 20 2.16 2.49
F3 4 1.67 1.48
F4 4 1.85 1.63
F5 15 2.03 2.73
F6 10 1.59 1.94
F7 7 2.08 2.30
F8 23 2.01 3.54
F9 5 2.02 2.14
F10 20 2.21 2.96
Figure 3.CPU time with KP instances

Table 5.Comparative CPU time with MKP instances
N° n M SBABO LBABO
1 6 10 3.09 2.99
2 10 10 3.38 2.72
3 15 10 3.49 3.30
4 20 10 3.55 4.28
5 28 10 3.75 5.11
6 39 5 4.97 5.45
7 50 5 5.95 5.83
Figure 4.CPU time with MKP instances.
The ABO algorithm is a new swarm optimization algorithm. Considering its young age, there are
few applications in optimization problems based on ABO algorithm. The main purpose of this
paper is to validate that the ABO method is also effective for binary combinatorial optimization
problems. That is why we proposed two discrete binary versions of ABO algorithm (called
SBABO and LBABO) which led to two efficient ABO algorithm versions to deal with binary
optimization problems.
During the different experiments, we noticed that SBABO algorithm explored the search space
better than LBABO (see Figure 5). This comes down to the use of the Sigmoid Function and
probability model to generate binary solutions. As shown in Figure 5, LBABO converges faster
than SBABO which explains its results with MKP instances. It is notable that the performance of
the algorithm is insensitive to their parameters such as lp1 and lp2. These two parameters
influence the good balance between exploration and exploitation of the search space. The
diversity of the proposed algorithms is assured by the use of the elitism selection which
guarantees that the best solutions are kept in each generation. The proposed algorithms can be
implemented easily for other binary optimization problems.

Figure 5. Evolution of best solution with KP and MKP using SBABO and LBABO
5. CONCLUSION AND PERSPECTIVES
In this paper, two discrete binary versions of African Buffalo Optimization algorithm are
proposed. This contribution has two-fold aims: the first aim is to propose a binary version of
ABO algorithm to deal with binary optimization problems. The second aim is to prove the
effectiveness of the ABO algorithm in solving NP-hard combinatorial optimization problems. In
the first version called SBABO we used the sigmoid function and probability model to generate
binary solutions. In the second version called LBABO we used some logical operator to operate
the binary solution. The proposed algorithms are used to solving two NP-hard binary
combinatorial optimization problems: KP and MKP problems. The obtained results are compared
with the harmony search algorithm (NGHS), the best known solution and the Binary Particle
Swarm Optimization algorithm (BPSO) which has a common point with the proposed SBABO
algorithm (the two algorithms used the sigmoid function). The experimental studies prove the
feasibility and the effectiveness of the proposed algorithms. They proved that the proposed
algorithms give good and promising results. However, there are several issues to improve the
proposed algorithms. Firstly, in order to improve the performance of the proposed algorithms, we
recommend integrating of a local search method in the algorithms core. In addition, hybridization
with other operations inspired by other popular algorithms such as Genetic algorithm, Particle
Swarm Optimization or Cuckoo Search will also be potentially fruitful. The proposed algorithms

can be also applied to solve many other binary optimization problems and real industrial
problems.
REFERENCES
[1] Odili J.B, Kahar M.N.M, Anwar S. African Buffalo Optimization: A Swarm-Intelligence Technique.
Procedia Computer Science 76. Elsevier, 2015. 443 – 448.
[2] Odili J.B, Kahar M.N.M. Solving the Traveling Salesman’s Problem Using the African Buffalo
Optimization. Computational Intelligence and Neuroscience.Volume 2016, Article ID 1510256.
Hindawi Publishing Corporation, 2015.
[3] Odili J, Kahar M. N. M, Noraziah A and Kamarulzaman. S. F. A comparative evaluation of swarm
intelligence techniques for solving combinatorial optimization problems. International Journal of
Advanced Robotic Systems. DOI: 10.1177/1729881417705969. 2017.
[4] Padmapriya R. Maheswari D. Channel Allocation Optimization using African Buffalo Optimization-
Super Vector Machine for Networks. Asian Journal of Information Technology, 2017. DOI:
10.3923/ajit.2017.783.788.16: 783-788.
[5] Liao C.J, Tseng C.T. and Luarn P. A discrete version of particle swarm optimization for flowshop
scheduling problems. Computers & Operations Research. Elsevier, 2007. Vol. 34, No. 10, pp.3099–
3111.
[6] Huang S. H, Tian N, Wang.Y and Ji Z. Multi-objective flexible job-shop scheduling problem using
modified discrete particle swarm optimization.Springer Plus, 2016. 5:1432.
[7] Ammi M, Chikhi S. Cooperative Parallel Metaheuristics based Penguin Optimization Search for
Solving the Vehicle Routing Problem. International Journal of Applied Metaheuristic Computing,
2016. Vol 7. Issue 1.
[8] Gherboudj, A. and Chikhi, S. BPSO algorithms for knapsack problem. In Özcan, A., Zizka, J. and
Nagamalai, D. (Eds.): WiMo/CoNeCo, CCIS, Springer. 2011. Vol. 162, pp.217–227.
[9] Kong M. and Tian P. Apply the particle swarm optimization to the multidimensional knapsack
problem. inRutkowski, L. et al. (Eds.): Proc. ICAISC 2006, LNAI, Springer.2006. Vol. 4029,
pp.1140–1149.
[10] Julstrom B.A. Greedy. Genetic, and greedy genetic algorithms for the quadratic knapsack problem. In
Proc. GECCO ‘05 Proceedings of the 2005 Conference on Genetic and Evolutionary Computation,
Publisher, ACM, 2005. pp.607–614.
[11] Singh A. and Baghel A.S. A new grouping genetic algorithm for the quadratic multiple knapsack
problem. In Cotta, C. and van Hemert, J. (Eds.): Proc. EvoCOP 2007, LNCS, Springer. 2007. Vol.
4446, pp.210–218.
[12] Pisinger D. Where are the hard knapsack problems? Computers and Operations Research, 2005. Vol.
32, No. 9, pp.2271–2284.
[13] Chu P.C, Beasley J.E. A genetic algorithm for the multidimensional knapsack problem. Journal of
Heuristics,1998. Vol. 4, No. 1, pp.63–86.

[14] Angelelli E, Mansini R. and Speranza M.G. Kernel search: a general heuristic for the multi-
dimensional knapsack problem. Computers & Operations Research, Elsevier, 2010. Vol. 37, No. 13,
pp.2017–2026.
[15] Vasquez M, Vimont Y. Improved results on the 0-1 multidimensional knapsack problem. European
Journal of Operational Research,2005. Vol. 165, No. 1, pp.70–81.
[16] Zou D, Gao L, Li S. and Wu J. Solving 0-1 knapsack problem by a novel global harmony search
algorithm. Applied Soft Computing, The Impact of Soft Computing for the Progress of Artificial
Intelligence, March, 2011. Vol. 11, No. 2, pp.1556–1564.
[17] Kennedy J, Eberhart R C. A discrete binary version of the particle swarm algorithm. In Proceedings
of the World Multiconference on Systemics, Cybernetics and Informatics, Piscatawary, NJ,1997.
pp.4104–4109.
AUTHOR
Amira Gherboudj is Senior Lecturer at the Algerian University of “Frères Mentouri,
Constantine 1”. Dr. Gherboudj received her PhD degree in computer science in 2013 from the
University of “AbdelhamidMehri, Constantine 2”.Her research interests include combinatorial
optimization methods and their applications to solve several problems from different domains.

TWO DISCRETE BINARY VERSIONS OF AFRICAN BUFFALO OPTIMIZATION METAHEURISTIC

More Related Content

What's hot (18)

Similar to TWO DISCRETE BINARY VERSIONS OF AFRICAN BUFFALO OPTIMIZATION METAHEURISTIC (20)

More from cscpconf (20)

Recently uploaded (20)

TWO DISCRETE BINARY VERSIONS OF AFRICAN BUFFALO OPTIMIZATION METAHEURISTIC