Mathematics,
Год журнала:
2021,
Номер
9(18), С. 2230 - 2230
Опубликована: Сен. 10, 2021
Magnetorheological
(MR)
dampers
play
a
crucial
role
in
various
engineering
systems,
and
how
to
identify
the
control
parameters
of
MR
damper
models
without
any
prior
knowledge
has
become
burning
problem.
In
this
study,
more
accurately,
an
improved
manta
ray
foraging
optimization
(IMRFO)
is
proposed.
The
new
algorithm
designs
searching
factor
according
weak
exploration
ability
MRFO,
which
can
effectively
increase
global
algorithm.
To
prevent
premature
convergence
local
optima,
adaptive
weight
coefficient
based
on
Levy
flight
designed.
Moreover,
by
introducing
Morlet
wavelet
mutation
strategy
algorithm,
space
adaptively
adjusted
enhance
step
out
stagnation
rate.
performance
IMRFO
evaluated
two
sets
benchmark
functions
results
confirm
competitiveness
proposed
Additionally,
applied
identifying
dampers,
simulation
reveal
effectiveness
practicality
applications.
Journal Of Big Data,
Год журнала:
2021,
Номер
8(1)
Опубликована: Янв. 26, 2021
Abstract
The
leading
approaches
in
Machine
Learning
are
notoriously
data-hungry.
Unfortunately,
many
application
domains
do
not
have
access
to
big
data
because
acquiring
involves
a
process
that
is
expensive
or
time-consuming.
This
has
triggered
serious
debate
both
the
industrial
and
academic
communities
calling
for
more
data-efficient
models
harness
power
of
artificial
learners
while
achieving
good
results
with
less
training
particular
human
supervision.
In
light
this
debate,
work
investigates
issue
algorithms’
hungriness.
First,
it
surveys
from
different
perspectives.
Then,
presents
comprehensive
review
existing
methods
systematizes
them
into
four
categories.
Specifically,
survey
covers
solution
strategies
handle
data-efficiency
by
(i)
using
non-supervised
algorithms
are,
nature,
data-efficient,
(ii)
creating
artificially
data,
(iii)
transferring
knowledge
rich-data
poor-data
domains,
(iv)
altering
data-hungry
reduce
their
dependency
upon
amount
samples,
way
they
can
perform
well
small
samples
regime.
Each
strategy
extensively
reviewed
discussed.
addition,
emphasis
put
on
how
interplay
each
other
order
motivate
exploration
robust
algorithms.
Finally,
delineates
limitations,
discusses
research
challenges,
suggests
future
opportunities
advance
machine
learning.
Swarm and Evolutionary Computation,
Год журнала:
2021,
Номер
67, С. 100973 - 100973
Опубликована: Авг. 20, 2021
Bio-inspired
optimization
(including
Evolutionary
Computation
and
Swarm
Intelligence)
is
a
growing
research
topic
with
many
competitive
bio-inspired
algorithms
being
proposed
every
year.
In
such
an
active
area,
preparing
successful
proposal
of
new
algorithm
not
easy
task.
Given
the
maturity
this
field,
proposing
technique
innovative
elements
no
longer
enough.
Apart
from
novelty,
results
reported
by
authors
should
be
proven
to
achieve
significant
advance
over
previous
outcomes
state
art.
Unfortunately,
all
proposals
deal
requirement
properly.
Some
them
fail
select
appropriate
benchmarks
or
reference
compare
with.
other
cases,
validation
process
carried
out
defined
in
principled
way
(or
even
done
at
all).
Consequently,
significance
presented
studies
cannot
guaranteed.
work
we
review
several
recommendations
literature
propose
methodological
guidelines
prepare
proposal,
taking
these
issues
into
account.
We
expect
useful
only
for
authors,
but
also
reviewers
editors
along
their
assessment
contributions
field.
Applied Sciences,
Год журнала:
2021,
Номер
11(14), С. 6449 - 6449
Опубликована: Июль 13, 2021
In
the
past
few
decades,
metaheuristics
have
demonstrated
their
suitability
in
addressing
complex
problems
over
different
domains.
This
success
drives
scientific
community
towards
definition
of
new
and
better-performing
heuristics
results
an
increased
interest
this
research
field.
Nevertheless,
studies
been
focused
on
developing
algorithms
without
providing
consolidation
existing
knowledge.
Furthermore,
absence
rigor
formalism
to
classify,
design,
develop
combinatorial
optimization
represents
a
challenge
field’s
progress.
study
discusses
main
concepts
challenges
area
proposes
code
metaheuristics.
We
believe
these
contributions
may
support
progress
field
increase
maturity
as
problem
solvers
analogous
other
machine
learning
algorithms.
Energy Reports,
Год журнала:
2021,
Номер
7, С. 1217 - 1233
Опубликована: Фев. 21, 2021
Due
to
the
intermittent,
fluctuating
and
random
characteristics
of
wind
system,
output
power
will
become
unstable
with
change
wind,
which
brings
severe
challenges
safe
stable
operation
system.
An
effective
way
solve
this
problem
is
accurately
forecast
speed.
This
paper
presents
a
novel
speed
combination
forecasting
model
based
on
decomposition.
The
innovation
as
follows.
(a)
In
view
speed,
variational
mode
decomposition
algorithm
introduced
decompose
historical
data
obtain
series
components
different
frequencies.
(b)
Echo
state
network
good
ability
selected
each
component.
(c)
To
that
performance
echo
greatly
affected
by
parameters
reservoir,
an
improved
whale
optimization
proposed
optimize
these
parameters.
optimized
improves
effect.
(d)
final
results
are
obtained
adding
values
(e)
developed
verified
using
two
actual
collected
sets
ultra-short-term
short-term
Compared
some
state-of-the-art
models,
comparison
result
curve
between
value
error
distribution,
histogram
indicators,
related
statistical
Taylor
diagram
show
has
higher
prediction
accuracy
able
reflect
laws
correctly.
