Third International Conference on Intelligent Computing and Human-Computer Interaction (ICHCI 2022),
Год журнала:
2023,
Номер
unknown
Опубликована: Янв. 13, 2023
Various
emergencies
occur
frequently,
posing
threats
and
challenges
to
people’s
lives
social
security.
In
consequence,
the
evacuation
of
multi-Agent
has
become
a
significant
part
emergency
response
process.
However,
few
existing
works
only
focus
on
small
number
agents,
which
does
not
consider
problem
cooperation
caused
by
increase
agents
impact
emergencies.
Therefore,
framework
for
event-driven
is
proposed
in
this
paper,
includes
three
parts:
event
collection,
sending,
task
execution.
During
execution,
are
divided
into
groups
select
leader
group,
while
other
group
move
with
leader.
Then,
reinforcement
learning
algorithm
Space
Multi-Agent
Deep
Deterministic
Policy
Gradient
(SMADDPG),
used
path
planning.
addition,
state,
action
reward
based
Markov
game
designed,
an
environment
presented
as
scenario.
The
experiment
results
show
that
method
can
shorten
length
path,
improve
interoperability
between
when
occur,
provide
decision-making
reference
departments
formulate
plans.
International Journal of Digital Earth,
Год журнала:
2023,
Номер
16(1), С. 691 - 714
Опубликована: Март 2, 2023
To
avoid
crowd
evacuation
simulations
depending
on
2D
environments
and
real
data,
we
propose
a
framework
for
modeling
simulation
by
applying
deep
reinforcement
learning
(DRL)
3D
physical
(3DPEs).
In
3DPEs,
construct
scenarios
from
the
aspects
of
geometry,
semantics
physics,
which
include
environment,
agents
their
interactions,
provide
training
samples
DRL.
DRL,
design
double
branch
feature
extraction
combined
actor
critic
network
as
DRL
policy
value
function
use
clipped
surrogate
objective
with
polynomial
decay
to
update
policy.
With
unified
configuration,
conduct
simulations.
one
exit,
reproduce
verify
bottleneck
effect
congested
crowds
explore
impact
exit
width
agent
characteristics
(number,
mass
height)
evacuation.
two
exits
uniform
(nonuniform)
distribution
agents,
(width
relative
position)
(height,
initial
location
distribution)
selection
Overall,
interactive
3DPEs
enable
adapt
different
simulate
laws
ISPRS International Journal of Geo-Information,
Год журнала:
2022,
Номер
11(4), С. 255 - 255
Опубликована: Апрель 13, 2022
At
present,
a
common
drawback
of
crowd
simulation
models
is
that
they
are
mainly
simulated
in
(abstract)
2D
environments,
which
limits
the
behaviors
observed
real
3D
environments.
Therefore,
we
propose
deep
reinforcement
learning-based
model
with
human-like
perceptron
and
policy
for
evacuation
environments
(HDRLM3D).
In
HDRLM3D,
vision-like
ray
(VLRP)
combine
it
redesigned
global
(or
local)
(GOLP)
to
form
perception
model.
We
double-branch
feature
extraction
decision
network
(DBFED-Net)
as
policy,
can
extract
features
make
behavioral
decisions.
Moreover,
validate
our
method’s
ability
reproduce
typical
phenomena
through
experiments
two
different
scenarios.
scenario
I,
bottleneck
effect
crowds
verify
effectiveness
advantages
HDRLM3D
by
comparing
classical
methods
terms
density
maps,
fundamental
diagrams,
times.
II,
agents’
navigation
obstacle
avoidance
demonstrate
unknown
other
trajectories
numbers
collisions.
PLoS ONE,
Год журнала:
2024,
Номер
19(1), С. e0293679 - e0293679
Опубликована: Янв. 18, 2024
Machine
learning
methods
and
agent-based
models
enable
the
optimization
of
operation
high-capacity
facilities.
In
this
paper,
we
propose
a
method
for
automatically
extracting
cleaning
pedestrian
traffic
detector
data
subsequent
calibration
ingress
model.
The
was
obtained
from
waiting
room
vaccination
center.
Walking
speed
distribution,
number
stops,
distribution
times,
locations
points
were
extracted.
Of
9
machine
algorithms,
random
forest
model
achieved
highest
accuracy
in
classifying
valid
noise.
proposed
microscopic
allows
more
accurate
capacity
assessment
testing,
procedural
changes
geometric
modifications
testing
parts
facility
adjacent
to
calibrated
parts.
results
show
that
achieves
state-of-the-art
performance
on
violent-flows
dataset.
has
potential
significantly
improve
efficiency
input
predictions
optimize
Agriculture,
Год журнала:
2023,
Номер
13(7), С. 1324 - 1324
Опубликована: Июнь 28, 2023
This
study
presents
a
multi-objective
optimization
method
for
tractor
driveline
based
on
the
diversity
maintenance
strategy
of
gradient
crowding.
The
objective
was
to
address
trade-off
between
high
power
and
low
fuel
consumption
rates
in
by
optimizing
distribution
ratios,
aiming
enhance
overall
driving
performance
reduce
consumption.
introduces
evaluating
crowding
non-inferior
solution
sets
during
selection
ensure
uniform
wide
solutions
while
maintaining
population
diversity.
transmission
ratio
is
optimized
varying
input
ratios
each
gear,
constraining
theoretical
rate,
common
ratio,
drive
adhesion
limit,
introducing
goal
loss
rate
specific
as
much
possible.
analysis
results
demonstrate
that
GC_NSGA-II
algorithm,
incorporating
evaluation
crowding,
achieves
greater
more
front
end.
After
verifying
showed
reduction
41.62
(±S.D.
0.44)%
62.8
0.56)%
consumption,
indicating
tractor’s
significantly
improved,
accompanied
substantial
rate.
These
findings
affirm
feasibility
proposed
provide
valuable
research
insights
enhancing
tractors.
