Biomimetics,
Год журнала:
2024,
Номер
10(1), С. 14 - 14
Опубликована: Дек. 31, 2024
The
RIME
algorithm
is
a
novel
physical-based
meta-heuristic
with
strong
ability
to
solve
global
optimization
problems
and
address
challenges
in
engineering
applications.
It
implements
exploration
exploitation
behaviors
by
constructing
rime-ice
growth
process.
However,
comes
couple
of
disadvantages:
limited
exploratory
capability,
slow
convergence,
inherent
asymmetry
between
exploitation.
An
improved
version
more
efficiency
adaptability
these
issues
now
the
form
Hybrid
Estimation
Rime-ice
Optimization,
short,
HERIME.
A
probabilistic
model-based
sampling
approach
estimated
distribution
utilized
enhance
quality
population
boost
its
capability.
roulette-based
fitness
distance
balanced
selection
strategy
used
strengthen
hard-rime
phase
effectively
balance
phases
We
validate
HERIME
using
41
functions
from
IEEE
CEC2017
CEC2022
test
suites
compare
accuracy,
stability
four
classical
recent
metaheuristic
algorithms
as
well
five
advanced
reveal
fact
that
proposed
outperforms
all
them.
Statistical
research
Friedman
Wilcoxon
rank
sum
also
confirms
excellent
performance.
Moreover,
ablation
experiments
effectiveness
each
individually.
Thus,
experimental
results
show
has
better
search
accuracy
effective
dealing
problems.
Remote Sensing,
Год журнала:
2025,
Номер
17(3), С. 500 - 500
Опубликована: Янв. 31, 2025
The
fragmentation
of
vessel
tracks
represents
a
significant
challenge
in
the
context
high-frequency
surface
wave
radar
(HFSWR)
tracking.
This
paper
proposes
new
track
segment
association
(TSA)
algorithm
that
integrates
optimal
tracklet
assignment,
iterative
discrimination,
and
multi-stage
association.
reformulates
assignment
task
as
an
state
search
problem
for
modeling
solution
purposes.
To
determine
whether
competing
old
tracklets
can
be
associated,
we
assume
existence
public
correlation
between
tracklets.
However,
due
to
fragmentation,
this
remains
unknown.
We
need
all
candidate
pairs
within
feasible
parameter
space,
using
fitness
function
value
evaluation
criterion.
match
are
considered
optimally
associated.
Since
process
involves
searching
across
multiple
dimensions,
it
constitutes
high-dimensional
optimization
problem.
accomplish
task,
catch
fish
(CFOA)
is
employed
its
ability
escape
local
optima
handle
optimization,
enhancing
reliability
assignment.
Furthermore,
achieve
precise
one-to-one
associations
by
assigning
through
method
proposed,
abbreviate
AN2O,
inverse
process,
which
assigns
tracklet,
abbreviated
AO2N.
dual
approach
further
complemented
discrimination
mechanism
evaluates
unselected
identify
potential
may
exist.
effectiveness
proposed
validated
field
experiment
data
from
HFSWR
Bohai
Sea
region,
demonstrating
capability
accurately
complex
data.
Biomimetics,
Год журнала:
2025,
Номер
10(3), С. 127 - 127
Опубликована: Фев. 20, 2025
The
Artificial
Gorilla
Troops
Optimizer
(GTO)
has
emerged
as
an
efficient
metaheuristic
technique
for
solving
complex
optimization
problems.
However,
the
conventional
GTO
algorithm
a
critical
limitation:
all
individuals,
regardless
of
their
roles,
utilize
identical
search
equations
and
perform
exploration
exploitation
sequentially.
This
uniform
approach
neglects
potential
benefits
labor
division,
consequently
restricting
algorithm’s
performance.
To
address
this
limitation,
we
propose
enhanced
Labor
Division
(LDGTO),
which
incorporates
natural
mechanisms
division
outcome
allocation.
In
phase,
stimulus-response
model
is
designed
to
differentiate
tasks,
enabling
gorilla
individuals
adaptively
adjust
based
on
environmental
changes.
allocation
three
behavioral
development
modes—self-enhancement,
competence
maintenance,
elimination—are
implemented,
corresponding
developmental
stages:
elite,
average,
underperforming
individuals.
performance
LDGTO
rigorously
evaluated
through
benchmark
test
suites,
comprising
12
unimodal,
25
multimodal,
10
combinatorial
functions,
well
two
real-world
engineering
applications,
including
four-bar
transplanter
mechanism
design
color
image
segmentation.
Experimental
results
demonstrate
that
consistently
outperforms
variants
seven
state-of-the-art
algorithms
in
most
cases.
Biomimetics,
Год журнала:
2025,
Номер
10(3), С. 153 - 153
Опубликована: Март 2, 2025
Intelligent
optimization
algorithms
are
crucial
for
solving
complex
engineering
problems.
The
Parrot
Optimization
(PO)
algorithm
shows
potential
but
has
issues
like
local-optimum
trapping
and
slow
convergence.
This
study
presents
the
Chaotic–Gaussian–Barycenter
(CGBPO),
a
modified
PO
algorithm.
CGBPO
addresses
these
problems
in
three
ways:
using
chaotic
logistic
mapping
random
initialization
to
boost
population
diversity,
applying
Gaussian
mutation
updated
individual
positions
avoid
premature
convergence,
integrating
barycenter
opposition-based
learning
strategy
during
iterations
expand
search
space.
Evaluated
on
CEC2017
CEC2022
benchmark
suites
against
seven
other
algorithms,
outperforms
them
convergence
speed,
solution
accuracy,
stability.
When
applied
two
practical
problems,
demonstrates
superior
adaptability
robustness.
In
an
indoor
visible
light
positioning
simulation,
CGBPO’s
estimated
closer
actual
ones
compared
PO,
with
best
coverage
smallest
average
error.
