Indian Journal of Science and Technology,
Год журнала:
2024,
Номер
17(45), С. 4722 - 4731
Опубликована: Дек. 14, 2024
Objectives:
To
evaluate
the
efficiency
of
task
prediction
and
resource
allocation
for
load
balancing
(LB)
in
cloud
environment
using
combined
approach
like
random
Forest(RF)
Particle
Swarm
optimization
Convolutional
Neural
Networks
(PSO-CNN)
allocation.
Methods:
The
ensemble
present
study
uses
Random
Forest
(RF),
a
machine
learning
(ML)
model
Optimization
(PSO+CNN),
bio-inspired
algorithm
Deep
Learning
(DL)
employs
PSO
techniques
to
optimize
CNN
order
address
investigation
algorithmic
DL.
results
show
that
suggested
outperforms
other
models
CNN-LSTM(Long
Short-term
memory),
CNN-GRU(Gated
Recurrent
Unit),
–SVM(Support
Vector
Machine)
increase
performance
efficacy
systems.
experiment
is
implemented
Python
assessed
Google
Cluster
dataset
accessible
public.
Findings:
use
ML
DL
are
found
be
more
efficient
infrastructure
than
conventional
methods.
examines
RF,
hybrid
RF-PSO-CNN
models.
accuracy,
precision,
F1.
Score
metrics
were
used
assess
classification
recommended
them
with
an
accuracy
90%
contrasted
methods
CNN-LSTM,
CNN-
GRU
PSO-SVM.
As
result,
both
assessment
consumption
proposed
performs
effectively.
Novelty:
novel
suggests
LB
Computing.
predicted
by
RF
assigned
chosen
CNN,
thereby
improving
Most
research
any
two
or
either
predicting
tasks
scheduled
which
allocate.
combination
(RF)
method,
(PSO)
(CNN)
concurrently
it
effectiveness
context.
Keywords:
Load
Balancing
(LB),
Task
scheduling,
Resource
allocation,
(CNN),
One
of
the
leading
diseases
globally
is
cancer
and
breast
not
exempted.
The
objective
WHO
Global
Breast
Cancer
Initiative
(GBCI)
to
reduce
global
mortality
by
2.5%
per
year,
thereby
averting
2.5
million
deaths
between
2020
2040.
three
pillars
toward
achieving
these
objectives
are:
health
promotion
for
early
detection;
timely
diagnosis;
comprehensive
management.
In
this
study
we
propose
an
detection
technique
in
combatting
diagnosis
combining
strength
both
PSO
(Particle
Swarm
Optimization)
BPSO
(Binary
Particle
achieve
optimal
solution.
results
obtained
indicated
superiority
Hybrid
PSO-BPSO
model
over
existing
solution
accuracy
98.82%
on
WBCD
WDBC
datasets.
Scientific Reports,
Год журнала:
2025,
Номер
15(1)
Опубликована: Март 18, 2025
The
problems
in
the
current
medical
equipment
measurement
work
have
made
many
institutions
begin
to
consider
establishing
a
convenient,
long-term
stable
and
low-cost
service
model
by
internal
standards,
but
they
not
further
in-depth
exploration
construction
of
perfect
standard
establish
feasibility
evaluation
system.
This
study
aims
construct
system
for
standards
based
on
Group
Decision
Making-Analytical
Hierarchy
Processes
provide
reference
basis
decision-making
establishment
equipment.
A
is
constructed,
which
includes
5
main
criteria-level
indicators
14
sub-criteria-level
indicators.
relative
weights
are
calculated
constructing
judgment
matrix
through
pairwise
comparisons
using
Saaty
scale
method.
Additionally,
sensitivity
analysis
constructed
conducted
perturbation
Then,
we
applied
eight
different
types
seven
institutions.
Differences
categories
both
an
impact
values
results
this
show
that,
can
transform
problem
exploring
into
multi-indicator
quantitative
problem,
making
difficult-to-quantify
process
more
scientific
objective.
Scientific Reports,
Год журнала:
2024,
Номер
14(1)
Опубликована: Ноя. 21, 2024
Swarm
Intelligence-based
metaheuristic
algorithms
are
widely
applied
to
global
optimization
and
engineering
design
problems.
However,
these
often
suffer
from
two
main
drawbacks:
susceptibility
the
local
optima
in
large
search
space
slow
convergence
rate.
To
address
issues,
this
paper
develops
a
novel
cooperative
algorithm
(CMA),
which
is
inspired
by
heterosis
theory.
Firstly,
simulating
hybrid
rice
(HRO)
constucted
based
on
theory,
population
sorted
fitness
divided
into
three
subpopulations,
corresponding
maintainer,
restorer,
sterile
line
HRO,
respectively,
engage
evolution.
Subsequently,
each
subpopulation,
three-phase
avoidance
technique-Search-Escape-Synchronize
(SES)
introduced.
In
phase,
well-established
Particle
Optimization
(PSO)
used
for
exploration.
During
escape
energy
dynamically
calculated
agent.
If
it
exceeds
threshold,
large-scale
Lévy
flight
jump
performed;
otherwise,
PSO
continues
conduct
search.
synchronize
best
solutions
subpopulations
shared
through
an
elite-based
strategy,
while
classical
Ant
Colony
employed
perform
fine-tuned
near
optimal
solutions.
This
process
accelerates
convergence,
maintains
diversity,
ensures
balanced
transition
between
exploration
exploitation.
validate
effectiveness
of
CMA,
study
evaluates
using
26
well-known
benchmark
functions
5
real-world
Experimental
results
demonstrate
that
CMA
outperforms
10
state-of-the-art
evaluated
study,
very
promising
problem
solving.
Mathematics,
Год журнала:
2024,
Номер
12(22), С. 3464 - 3464
Опубликована: Ноя. 6, 2024
Metaheuristic
algorithms
(MAs)
now
are
the
standard
in
engineering
optimization.
Progress
computing
power
has
favored
development
of
new
MAs
and
improved
versions
existing
methods
hybrid
MAs.
However,
most
(especially
algorithms)
have
very
complicated
formulations.
The
present
study
demonstrated
that
it
is
possible
to
build
a
simple
metaheuristic
algorithm
combining
basic
classical
MAs,
including
modifications
optimization
formulation
maximize
computational
efficiency.
(SHGWJA)
developed
here
combines
two
methods,
namely
grey
wolf
optimizer
(GWO)
JAYA,
widely
used
problems
continue
attract
attention
scientific
community.
SHGWJA
overcame
limitations
GWO
JAYA
exploitation
phase
using
elitist
strategies.
proposed
was
tested
successfully
seven
“real-world”
taken
from
various
fields,
such
as
civil
engineering,
aeronautical
mechanical
(included
CEC
2020
test
suite
on
real-world
constrained
problems)
robotics;
these
include
up
14
variables
721
nonlinear
constraints.
Two
representative
mathematical
(i.e.,
Rosenbrock
Rastrigin
functions)
1000
were
also
solved.
Remarkably,
always
outperformed
or
competitive
with
other
state-of-the-art
competition
winners
high-performance
all
cases.
In
fact,
found
global
optimum
best
cost
at
0.0121%
larger
than
target
optimum.
Furthermore,
robust:
(i)
cases,
obtained
0
near-0
deviation
runs
practically
converged
solution;
(ii)
optimized
0.0876%
design;
(iii)
function
evaluations
35%
average
cost.
Last,
ranked
1st
2nd
for
speed
its
fastest
highly
their
counterpart
recorded
