Page 195 - 2024-Vol20-Issue2

191 |
ture. Many algorithms have been invented to emulate the
natural systems in such a way that serves the industrial and
scientific kinds of problems. The following are the main ex-
amples of these algorithms. ant colony (AC) algorithm [2]
was proposed in 1992 as an algorithm that follows the ant
colony intelligence of communication in solving the optimiza-
tion problems. In this algorithm, the searching efficiency was
improved by increasing the randomness to present variety in
the solutions and avoiding falling in local optima [3]. In 1995,
particle swarm optimization (PSO) [4] was proposed. This
algorithm is based on migration movement of a swarm of
birds over a wide searching area. In 2010, an algorithm called
cuckoo search algorithm (CS) [5] that relies on the cuckoo
species co-parsing in parasitism was presented. An algorithm
was proposed in 2012 known as flower pollination algorithm
(FPA) [6]. This algorithm is based on the different pollination
methods used in agricultural fields and orchards. In addition,
there are some algorithms that were inspired from the natu-
ral behavior of elements via their chemical reactions without
the presence of catalyst such as Multi objective Atomic Or-
bital Search (MOAOS) [7]. Furthermore, some algorithms
were created based on the physical changes such as big bang-
big crunch algorithm [8], central force optimization [9] and
An improved pedestrian dead reckoning algorithm based on
smartphone [10].
Generally, there are many classifications for the optimization
algorithms. However, in this work, the classification that is
relied on follows the source of inspiration as in [1] at which
the algorithms are divided into four categories:
1) Swarm intelligence (SI) based
2) Bio-inspired, but not SI based
3) Physics and chemistry based
4) Other algorithms
The SI algorithms are the most popular type of algorithms.
All SI algorithms use multi–agent or multi-particle. There
are many reasons for this popularity. One of the most impor-
tant reasons is the information- sharing between particles or
agents which results in faster convergence to the optimum
situation. Another reason for the popularity of SI algorithms
is that it can be parallelized easily, so it can easily be worked followed in presenting the proposed algorithm. Therefore,
in large scale of optimization. There are many examples of
this type algorithms such as Modified camel travelling be-
havior algorithm (MCA) [11], Artificial bee colony (ABC)
[12], Bat algorithm (BA) [13], Bee colony optimization(BCO)
[14], and Virtual ant algorithm (VA) [15]. The second type
of this classification is the bio-inspired that are not SI-based
algorithms. These algorithms depend on the behavior of a
certain organism or creature regardless its communication
with other mates. Fungi kingdom expansion algorithm [16],
biogeography-based optimization [17], brain storm optimiza-
tion [18], fish-school search [19], and shuffled frog leaping
Al-mtory, Alnahwi & Ali
algorithm [20] are examples of this type of algorithms. The
third inspiration comes from the physical and chemical prop-
erties. The physical algorithms are inspired from the physical
theories and laws, while the chemical algorithms are inspired
from the interactions and the chemical properties of materials.
There are many algorithms that are physically and chemically
inspired such as black hole [21], charge system search [22],
and water cycle algorithm [23]. The algorithm introduces in
this paper which is called Corrosion Diffusion Optimization
Algorithm (CDOA) is categorized in this type of classifica-
tion. Finally, the last type of this classification is called other
algorithms. Sometimes, the researchers introduce and develop
algorithms that are difficult to include under the above three
types. It may be an algorithm inspired from nature but it is
not bio, physically, or chemically. Therefore, it is placed un-
der the above mentioned category. Examples of this type are
anarchic society optimization [24] and artificial cooperative
search [25]. Variety of optimization algorithms are proposed
due to several reasons such as the diversity and multiplicity of
problem in engineer and industrial applications, and there is
no algorithm that solves all engineering and industrial appli-
cations problems according to No Free Lunch (NFL) theorem
[26].
In this paper, a new algorithm called Corrosion Diffusion
Optimization Algorithm (CDOA) is presented. It is based on
the corrosion distribution on the metal surfaces. This paper is
organized as follows. Section II discusses the method from
which the algorithm is inspired. The implementation of the
proposed algorithm is demonstrated in Section III, while Sec-
tion IV verifies the performance of the proposed algorithm
with the aid of some benchmark functions to compare it with
some other well-known algorithms. Some engineering appli-
cations are presented in Section V to fortify the superiority
of this algorithm. Section VI is the final conclusion of this
research.
II. BEHAVIOR OF CORROSION DIFFUSION
This section demonstrates the inspiration method that was
this section is divided into the following two subsections.
A. Inspiration Method
Corrosion can be defined as a failure or dissolution of a metal
or alloy due to a chemical or electrochemical interaction with
the medium surrounding it. When the corrosion is caused by
physical reasons, it is called wear. However, the corrosion that
results from chemical factors and with the help of mechani-
cal factors takes other names such as erosion corrosion and
retting corrosion[27], respectively. The study of corrosion is
an important matter for humanity because of its impact on the
everyday life. The corrosion may results in a huge financial