Metaheuristic
techniques
are
capable
of
representing
optimization
frames
with
their
specific
theories
as
well
objective
functions
owing
to
being
adjustable
and
effective
in
various
applications.
Through
the
deep
learning
models,
metaheuristic
algorithms
inspired
by
nature,
imitating
behavior
living
non-living
beings,
have
been
used
for
about
four
decades
solve
challenging,
complex,
chaotic
problems.
These
can
be
categorized
evolution-based,
swarm-based,
nature-based,
human-based,
hybrid,
or
chaos-based.
Chaos
theory,
a
useful
approach
understanding
neural
network
optimization,
has
basic
idea
viewing
dynamical
system
which
equation
schemes
utilized
from
space
pertaining
learnable
parameters,
namely
trajectory,
itself,
enables
description
evolution
training
behavior,
is
say
number
iterations
over
time.
The
examination
recent
studies
reveals
importance
chaos
sensitive
initial
conditions
randomness
properties
that
principally
emerging
on
complex
multimodal
landscape.
Chaotic
this
regard,
accelerates
speed
algorithm
while
also
enhancing
variety
movement
patterns.
significance
hybrid
developed
through
applications
different
domains
concerning
real-world
phenomena
well-known
benchmark
problems
literature
evident.
applied
networks
(DNNs),
branch
machine
learning.
In
respect,
features
DNNs
extensive
use
overviewed
explained.
Accordingly,
current
review
aims
at
providing
new
insights
into
deal
algorithms,
hybrid-based
metaheuristics,
chaos-based
metaheuristics
besides
presenting
information
development
essence
science
opportunities,
applicability-based
aspects
generation
well-informed
decisions.
Entropy,
Год журнала:
2021,
Номер
23(12), С. 1637 - 1637
Опубликована: Дек. 6, 2021
Moth-flame
optimization
(MFO)
algorithm
inspired
by
the
transverse
orientation
of
moths
toward
light
source
is
an
effective
approach
to
solve
global
problems.
However,
MFO
suffers
from
issues
such
as
premature
convergence,
low
population
diversity,
local
optima
entrapment,
and
imbalance
between
exploration
exploitation.
In
this
study,
therefore,
improved
moth-flame
(I-MFO)
proposed
cope
with
canonical
MFO's
locating
trapped
in
optimum
via
defining
memory
for
each
moth.
The
tend
escape
taking
advantage
adapted
wandering
around
search
(AWAS)
strategy.
efficiency
I-MFO
evaluated
CEC
2018
benchmark
functions
compared
against
other
well-known
metaheuristic
algorithms.
Moreover,
obtained
results
are
statistically
analyzed
Friedman
test
on
30,
50,
100
dimensions.
Finally,
ability
find
best
optimal
solutions
mechanical
engineering
problems
three
latest
test-suite
2020.
experimental
statistical
demonstrate
that
significantly
superior
contender
algorithms
it
successfully
upgrades
shortcomings
MFO.
Processes,
Год журнала:
2021,
Номер
9(12), С. 2276 - 2276
Опубликована: Дек. 18, 2021
Moth–flame
optimization
(MFO)
is
a
prominent
swarm
intelligence
algorithm
that
demonstrates
sufficient
efficiency
in
tackling
various
tasks.
However,
MFO
cannot
provide
competitive
results
for
complex
problems.
The
sinks
into
the
local
optimum
due
to
rapid
dropping
of
population
diversity
and
poor
exploration.
Hence,
this
article,
migration-based
moth–flame
(M-MFO)
proposed
address
mentioned
issues.
In
M-MFO,
main
focus
on
improving
position
unlucky
moths
by
migrating
them
stochastically
early
iterations
using
random
migration
(RM)
operator,
maintaining
solution
diversification
storing
new
qualified
solutions
separately
guiding
archive,
and,
finally,
exploiting
around
positions
saved
archive
guided
(GM)
operator.
dimensionally
aware
switch
between
these
two
operators
guarantees
convergence
toward
promising
zones.
M-MFO
was
evaluated
CEC
2018
benchmark
suite
dimension
30
compared
against
seven
well-known
variants
MFO,
including
LMFO,
WCMFO,
CMFO,
CLSGMFO,
LGCMFO,
SMFO,
ODSFMFO.
Then,
top
four
latest
high-performing
were
considered
experiments
with
different
dimensions,
30,
50,
100.
experimental
evaluations
proved
provides
exploration
ability
maintenance
employing
strategy
archive.
addition,
statistical
analyzed
Friedman
test
performance
contender
algorithms
used
experiments.
Alexandria Engineering Journal,
Год журнала:
2022,
Номер
64, С. 365 - 389
Опубликована: Сен. 22, 2022
This
paper
proposes
a
hybridized
version
of
the
Harris
Hawks
Optimizer
(HHO)
with
adaptive-hill-climbing
optimizer
to
tackle
economic
load
dispatch
(ELD)
problems.
ELD
is
an
important
problem
in
power
systems
that
tackled
by
finding
optimal
schedule
generation
units
minimize
fuel
conceptions
under
set
constraints.
Due
complexity
search
space,
as
it
rigid
and
deep,
exploitation
HHO
improved
hybridizing
recent
local
method
called
adaptive-hill
climbing.
The
can
navigate
several
potential
space
regions,
while
climbing
used
deeply
for
solution
each
region.
To
evaluate
proposed
approach,
six
versions
cases
various
complexities
constraints
have
been
which
are
6
1263
MW
demand,
13
1800
2520
15
2630
40
10500
140
49342
demand.
Furthermore,
algorithm
evaluated
on
two
real-world
units-1263
15units-2630
MW.
results
show
achieve
significant
performance
majority
experimented
cases.
It
best-reported
case
when
compared
well-established
methods.
Additionally,
obtains
second-best
10
In
conclusion,
be
alternative
solve
problems
efficient.
IEEE Access,
Год журнала:
2022,
Номер
10, С. 19254 - 19283
Опубликована: Янв. 1, 2022
This
paper
presents
a
new
hybrid
metaheuristic
algorithm,
the
Harris
Hawks
Optimizer-Arithmetic
Optimization
Algorithm
(hHHO-AOA),
as
we
have
named
it.
It
is
proposed
for
sizing
optimization
and
design
of
autonomous
microgrids.
The
algorithm
has
been
developed
based
on
operating
Optimizer
(HHO)
Arithmetic
(AOA)
in
uniquely
cooperative
manner.
expected
to
increase
solution
accuracy
by
increasing
diversity
during
an
process.
performance
verified
with
evaluation
metrics
well-known
statistical
tests.
According
Friedman
ranking
test,
performs
77.9%
better
than
HHO
78.6%
AOA.
Similarly,
checked
Wilcoxon
signed-rank
test
revealed
significant
superiority
compared
AOA
alone.
Later,
tested
microgrid
that
consists
photovoltaic
(PV)
system,
wind
turbine
(WT)
battery
energy
storage
system
(BESS),
diesel
generators
(DGs),
commercial
type
load.
For
optimal
capacity
planning
these
components,
problem
which
loss
power
supply
probability
(LPSP)
cost
(COE)
are
defined
objective
function
formulated.
done
produced
lowest
LPSP
COE
along
highest
rate
renewable
fraction
(RF).
In
conclusion,
it
demonstrated
hHHO-AOA
proved
itself
designing
reliable,
economical,
eco-friendly
microgrids
best
way.
Mathematics,
Год журнала:
2023,
Номер
11(10), С. 2340 - 2340
Опубликована: Май 17, 2023
In
this
study,
a
new
hybrid
metaheuristic
algorithm
named
Chaotic
Sand
Cat
Swarm
Optimization
(CSCSO)
is
proposed
for
constrained
and
complex
optimization
problems.
This
combines
the
features
of
recently
introduced
SCSO
with
concept
chaos.
The
basic
aim
to
integrate
chaos
feature
non-recurring
locations
into
SCSO’s
core
search
process
improve
global
performance
convergence
behavior.
Thus,
randomness
in
can
be
replaced
by
chaotic
map
due
similar
better
statistical
dynamic
properties.
addition
these
advantages,
low
consistency,
local
optimum
trap,
inefficiency
search,
population
diversity
issues
are
also
provided.
CSCSO,
several
maps
implemented
more
efficient
behavior
exploration
exploitation
phases.
Experiments
conducted
on
wide
variety
well-known
test
functions
increase
reliability
results,
as
well
real-world
was
applied
total
39
multidisciplinary
It
found
76.3%
responses
compared
best-developed
variant
other
chaotic-based
metaheuristics
tested.
extensive
experiment
indicates
that
CSCSO
excels
providing
acceptable
results.
