Research Square (Research Square),
Год журнала:
2024,
Номер
unknown
Опубликована: Март 6, 2024
Abstract
The
randomicity
and
fluctuation
of
the
wind
speed
will
influence
precision
forecast.
To
improve
forecast,
this
paper
presents
a
new
method
combined
forecast
based
on
second
decomposition
weighted
average.
First,
ICEEMDAN
is
used
to
get
different
sub-sequences,
then
fuzzy
entropy
judge
degree
confusion
sub-sequences.
In
paper,
ARIMA
model
predict
minimum
entropy.
And
other
subsequences
are
decomposed
by
BPNN,
VMD
predicted
NAR
BP
neural
network
with
suitable
weighting
ratio
for
average
PSO-LSTM
respectively,
ultimately
all
values
superimposed
final
prediction.
Experiments
were
conducted
using
three
datasets
eight
comparison
models
verify
validity
model.
prediction
analysis
was
carried
out
actual
measured
data
farm
in
Inner
Mongolia,
results
indicated
that
(1)
can
effectively
precision;
(2)
accuracy
secondary
greatly
improved
more
reliable;
(3)
Decompose
one
VMD,
it
network,
choose
appropriate
weight
achieve
better
results;
(4)
root
mean
square
error
(RMSE)
hybrid
1
0.28777,
0.22786
0.17128,
which
lower
than
models.
So,
workable
use
Bioengineering,
Год журнала:
2023,
Номер
10(7), С. 825 - 825
Опубликована: Июль 11, 2023
Early
detection
of
breast
lesions
and
distinguishing
between
malignant
benign
are
critical
for
cancer
(BC)
prognosis.
Breast
ultrasonography
(BU)
is
an
important
radiological
imaging
modality
the
diagnosis
BC.
This
study
proposes
a
BU
image-based
framework
BC
in
women.
Various
pre-trained
networks
used
to
extract
deep
features
images.
Ten
wrapper-based
optimization
algorithms,
including
marine
predator
algorithm,
generalized
normal
distribution
optimization,
slime
mold
equilibrium
optimizer
(EO),
manta-ray
foraging
atom
search
Harris
hawks
Henry
gas
solubility
path
finder
poor
rich
were
employed
compute
optimal
subset
using
support
vector
machine
classifier.
Furthermore,
network
selection
algorithm
was
determine
best
network.
An
online
dataset
test
proposed
framework.
After
comprehensive
testing
analysis,
it
found
that
EO
produced
highest
classification
rate
each
model.
It
accuracy
96.79%,
trained
only
feature
with
size
562
ResNet-50
Similarly,
Inception-ResNet-v2
had
second
96.15%
algorithm.
Moreover,
results
compared
those
literature.
2022 14th International Conference on Electronics, Computers and Artificial Intelligence (ECAI),
Год журнала:
2023,
Номер
unknown, С. 1 - 6
Опубликована: Июнь 29, 2023
This
paper
presents
the
collision-free
path
planning
approach
for
automated
guided
vehicle
(AGV)
in
an
intelligent
warehouse
environment,
optimized
by
means
of
recent
well-known
meta-heuristic
algorithms.
novel
is
assessment
and
possibilities
scheduling
multi-AGVs
to
complete
given
tasks
a
minimal
travel
distance
with
optimal
operation
time.
Six
different
metaheuristic
algorithms
such
as
PSO,
MELGWO,
GTO,
SFS,
MPA
chaotic-improved
are
compared
AGVs
optimization
capability.
In
order
test
robustness
proposed
approaches,
four
scenarios
presented
which
include
general
obstacle
avoidance
three
simple
maps
that
treated
environment.
each
scenario,
obstacles
placed
way
increase
overall
complexity
AGV
reach
target
destination.
The
exploration
exploitation
phases
algorithm
enhanced
simultaneously
replacing
conventional
Gaussian
random
chaotic
operators
ensure
its
effectiveness
optimization.
outperforms
other
based
on
statistical
analysis
results
improvement
11.0171%
comparison
unoptimized
probabilistic
roadmap
method
(PRM)
planner.
conclusion,
can
be
efficiently
all
aforementioned
environments.
2021 16th International Conference on Intelligent Systems and Knowledge Engineering (ISKE),
Год журнала:
2023,
Номер
unknown, С. 615 - 619
Опубликована: Ноя. 17, 2023
Aiming
at
the
issue
of
power
grid
substation
location
planning
in
grid,
a
model
is
established
with
goal
economy,
and
based
on
Multi-strategy
Improved
Marine
Predators
Algorithm
(MIMPA)
proposed
to
solve
model.
The
algorithm
introduces
Sobol
sequence
low-difference
make
initial
site
randomly
evenly
distributed
solution
space,
which
ensures
ergodicity
diversity
compared
random
sequence.
Differential
Evolution
(DE)
used
obtain
optimal
each
generation
adopts
mutation,
crossover,
selection
problem
that
(MPA)
difficult
jump
out
local
solution,
thus
providing
excellent
candidates
for
final
decision.
tested
through
an
example
differential
evolution
Firefly
(FA),
verifies
superiority,
feasibility
practicability
algorithm.
Research Square (Research Square),
Год журнала:
2024,
Номер
unknown
Опубликована: Март 6, 2024
Abstract
The
randomicity
and
fluctuation
of
the
wind
speed
will
influence
precision
forecast.
To
improve
forecast,
this
paper
presents
a
new
method
combined
forecast
based
on
second
decomposition
weighted
average.
First,
ICEEMDAN
is
used
to
get
different
sub-sequences,
then
fuzzy
entropy
judge
degree
confusion
sub-sequences.
In
paper,
ARIMA
model
predict
minimum
entropy.
And
other
subsequences
are
decomposed
by
BPNN,
VMD
predicted
NAR
BP
neural
network
with
suitable
weighting
ratio
for
average
PSO-LSTM
respectively,
ultimately
all
values
superimposed
final
prediction.
Experiments
were
conducted
using
three
datasets
eight
comparison
models
verify
validity
model.
prediction
analysis
was
carried
out
actual
measured
data
farm
in
Inner
Mongolia,
results
indicated
that
(1)
can
effectively
precision;
(2)
accuracy
secondary
greatly
improved
more
reliable;
(3)
Decompose
one
VMD,
it
network,
choose
appropriate
weight
achieve
better
results;
(4)
root
mean
square
error
(RMSE)
hybrid
1
0.28777,
0.22786
0.17128,
which
lower
than
models.
So,
workable
use
