Mathematics,
Год журнала:
2024,
Номер
12(19), С. 3140 - 3140
Опубликована: Окт. 7, 2024
In
urban
logistics,
effective
maintenance
is
crucial
for
maintaining
the
reliability
and
efficiency
of
energy
supply
systems,
impacting
both
asset
performance
operational
stability.
This
paper
addresses
scheduling
routing
plans
power
generation
assets
over
a
multi-period
horizon.
We
model
this
problem
as
team
orienteering
problem.
To
address
challenge,
we
propose
dual
approach:
novel
reinforcement
learning
(RL)
framework
biased-randomized
heuristic
algorithm.
The
RL-based
method
dynamically
learns
from
real-time
data
evolving
conditions,
adapting
to
changes
in
health
failure
probabilities
optimize
decision
making.
addition,
develop
apply
algorithm
designed
provide
solutions
within
practical
computational
limits.
Our
approach
validated
through
series
experiments
comparing
RL
results
demonstrate
that,
when
properly
trained,
able
offer
equivalent
or
even
superior
compared
IET Collaborative Intelligent Manufacturing,
Год журнала:
2023,
Номер
5(1)
Опубликована: Март 1, 2023
Abstract
An
efficient
manufacturing
system
is
key
to
maintaining
a
healthy
economy
today.
With
the
rapid
development
of
science
and
technology
progress
human
society,
modern
becoming
increasingly
complex,
posing
new
challenges
both
academia
industry.
Ever
since
beginning
industrialisation,
leaps
in
have
always
accompanied
technological
breakthroughs
from
other
fields,
for
example,
mechanics,
physics,
computational
science.
Recently,
machine
learning
(ML)
technology,
one
crucial
subjects
artificial
intelligence,
has
made
remarkable
many
areas.
This
study
thoroughly
reviews
how
ML,
specifically
deep
(reinforcement)
learning,
motivates
ideas
addressing
challenging
problems
systems.
We
collect
literature
targeting
three
aspects:
scheduling,
packing,
routing,
which
correspond
pivotal
cooperative
production
links
today's
system,
that
is,
production,
logistics
respectively.
For
each
aspect,
we
first
present
discuss
state‐of‐the‐art
research.
Then
summarise
analyse
trends
point
out
future
research
opportunities
challenges.
IEEE Transactions on Intelligent Transportation Systems,
Год журнала:
2024,
Номер
25(6), С. 4754 - 4772
Опубликована: Янв. 2, 2024
This
paper
provides
a
systematic
overview
of
machine
learning
methods
applied
to
solve
NP-hard
Vehicle
Routing
Problems
(VRPs).
Recently,
there
has
been
great
interest
from
both
the
and
operations
research
communities
in
solving
VRPs
either
through
pure
or
by
combining
them
with
traditional
handcrafted
heuristics.
We
present
taxonomy
studies
on
paradigms,
solution
structures,
underlying
models,
algorithms.
Detailed
results
state-of-the-art
are
presented,
demonstrating
their
competitiveness
approaches.
The
survey
highlights
advantages
learning-based
models
that
aim
exploit
symmetry
VRP
solutions.
outlines
future
directions
incorporate
solutions
address
challenges
modern
transportation
systems.
IEEE Transactions on Intelligent Transportation Systems,
Год журнала:
2024,
Номер
25(8), С. 9253 - 9267
Опубликована: Авг. 1, 2024
Crowdshipping
problem
can
be
challenging
as
the
platform
are
continuously
but
sporadically
receiving
crowdshippers
and
delivery
tasks
with
heterogeneous
origin
destination.
In
this
paper,
dynamic
multi-vehicle
pickup
(DMV-PDPC)
is
considered.
Leveraging
deep
reinforcement
learning
framework,
attention
model
centralized
vehicle
network
(AMCVN)
method
developed.
Unlike
traditional
heuristic
or
existing
vehicle-changing
methods,
AMCVN
integrates
a
(CVN)
that
observe
state
information
of
all
vehicles,
enhancing
its
overall
performance.
each
decision-making
step,
CVN
monitors
vehicles
selects
one
vehicles.
Subsequently,
attention-based
route
generating
(RGN)
determines
next
node
to
visited
by
chosen
vehicle.
Instead
using
penalty
term
in
reward
function
regulate
sequence
visits
nodes,
more
precise
control
method,
namely
rolling
mask
scheme
(RMS),
implemented.
The
method's
evaluation
carried
out
via
simulation
experiment
real-world
road
network.
This
demonstrates
proposed
effectively
tackles
DMV-PDPC
challenge,
outperforming
current
state-of-the-art
learning-based
models
methods.
Moreover,
shows
exceptional
generalization
capabilities,
evidenced
adaptability
different
numbers
IEEE Transactions on Intelligent Transportation Systems,
Год журнала:
2023,
Номер
24(12), С. 15652 - 15666
Опубликована: Март 16, 2023
Airport
ground
handling
(AGH)
offers
necessary
operations
to
flights
during
their
turnarounds
and
is
of
great
importance
the
efficiency
airport
management
economics
aviation.
Such
a
problem
involves
interplay
among
that
leads
NP-hard
problems
with
complex
constraints.
Hence,
existing
methods
for
AGH
are
usually
designed
massive
domain
knowledge
but
still
fail
yield
high-quality
solutions
efficiently.
In
this
paper,
we
aim
enhance
solution
quality
computation
solving
AGH.
Particularly,
first
model
as
multiple-fleet
vehicle
routing
(VRP)
miscellaneous
constraints
including
precedence,
time
windows,
capacity.
Then
propose
construction
framework
decomposes
into
sub-problems
(i.e.,
VRPs)
in
fleets
present
neural
method
construct
these
sub-problems.
specific,
resort
deep
learning
parameterize
heuristic
policy
an
attention-based
network
trained
reinforcement
learning,
which
shared
across
all
Extensive
experiments
demonstrate
our
significantly
outperforms
classic
meta-heuristics,
heuristics
specialized
Besides,
empirically
verify
generalizes
well
instances
large
numbers
or
varying
parameters,
can
be
readily
adapted
solve
real-time
stochastic
flight
arrivals.
Our
code
publicly
available
at:
https://github.com/RoyalSkye/AGH.
