Artificial Intelligence Review,
Journal Year:
2025,
Volume and Issue:
58(3)
Published: Jan. 6, 2025
The
advent
of
the
intelligent
information
era
has
witnessed
a
proliferation
complex
optimization
problems
across
various
disciplines.
Although
existing
meta-heuristic
algorithms
have
demonstrated
efficacy
in
many
scenarios,
they
still
struggle
with
certain
challenges
such
as
premature
convergence,
insufficient
exploration,
and
lack
robustness
high-dimensional,
nonconvex
search
spaces.
These
limitations
underscore
need
for
novel
techniques
that
can
better
balance
exploration
exploitation
while
maintaining
computational
efficiency.
In
response
to
this
need,
we
propose
Artificial
Lemming
Algorithm
(ALA),
bio-inspired
metaheuristic
mathematically
models
four
distinct
behaviors
lemmings
nature:
long-distance
migration,
digging
holes,
foraging,
evading
predators.
Specifically,
migration
burrow
are
dedicated
highly
exploring
domain,
whereas
foraging
predators
provide
during
process.
addition,
ALA
incorporates
an
energy-decreasing
mechanism
enables
dynamic
adjustments
between
exploitation,
thereby
enhancing
its
ability
evade
local
optima
converge
global
solutions
more
robustly.
To
thoroughly
verify
effectiveness
proposed
method,
is
compared
17
other
state-of-the-art
on
IEEE
CEC2017
benchmark
test
suite
CEC2022
suite.
experimental
results
indicate
reliable
comprehensive
performance
achieve
superior
solution
accuracy,
convergence
speed,
stability
most
cases.
For
29
10-,
30-,
50-,
100-dimensional
functions,
obtains
lowest
Friedman
average
ranking
values
among
all
competitor
methods,
which
1.7241,
2.1034,
2.7241,
2.9310,
respectively,
12
again
wins
optimal
2.1667.
Finally,
further
evaluate
applicability,
implemented
address
series
cases,
including
constrained
engineering
design,
photovoltaic
(PV)
model
parameter
identification,
fractional-order
proportional-differential-integral
(FOPID)
controller
gain
tuning.
Our
findings
highlight
competitive
edge
potential
real-world
applications.
source
code
publicly
available
at
https://github.com/StevenShaw98/Artificial-Lemming-Algorithm
.
Artificial Intelligence Review,
Journal Year:
2025,
Volume and Issue:
58(3)
Published: Jan. 6, 2025
Abstract
Numerical
optimization
and
point
cloud
registration
are
critical
research
topics
in
the
field
of
artificial
intelligence.
The
differential
evolution
algorithm
is
an
effective
approach
to
address
these
problems,
LSHADE-SPACMA,
winning
CEC2017,
a
competitive
variant.
However,
LSHADE-SPACMA’s
local
exploitation
capability
can
sometimes
be
insufficient
when
handling
challenges.
Therefore,
this
work,
we
propose
modified
version
LSHADE-SPACMA
(mLSHADE-SPACMA)
for
numerical
registration.
Compared
original
approach,
work
presents
three
main
innovations.
First,
present
precise
elimination
generation
mechanism
enhance
algorithm’s
ability.
Second,
introduce
mutation
strategy
based
on
semi-parametric
adaptive
rank-based
selective
pressure,
which
improves
evolutionary
direction.
Third,
elite-based
external
archiving
mechanism,
ensures
diversity
population
accelerate
convergence
progress.
Additionally,
utilize
CEC2014
(Dim
=
10,
30,
50,
100)
CEC2017
test
suites
experiments,
comparing
our
against:
(1)
10
recent
CEC
winner
algorithms,
including
LSHADE,
EBOwithCMAR,
jSO,
LSHADE-cnEpSin,
HSES,
LSHADE-RSP,
ELSHADE-SPACMA,
EA4eig,
L-SRTDE,
LSHADE-SPACMA;
(2)
4
advanced
variants:
APSM-jSO,
LensOBLDE,
ACD-DE,
MIDE.
results
Wilcoxon
signed-rank
Friedman
mean
rank
demonstrate
that
mLSHADE-SPACMA
not
only
outperforms
but
also
surpasses
other
high-performance
optimizers,
except
it
inferior
L-SRTDE
CEC2017.
Finally,
25
cases
from
Fast
Global
Registration
dataset
applied
simulation
analysis
potential
developed
technique
solving
practical
problems.
code
available
at
https://github.com/ShengweiFu?tab=repositories
https://ww2.mathworks.cn/matlabcentral/fileexchange/my-file-exchange
