How to Remove Document Management Hurdles with X-Docs?
Could Digital DIY Break Manufacturing Hierarchies?
1. Could
Digital
DIY
Break
Manufacturing
Hierarchies?
Ruth
Meyer,
Magnus
Josefsson
Centre
for
Policy
Modelling,
Manchester
Metropolitan
University
www.didiy.eu
2. What
is
Digital
DIY?
• Digital
DIY
=
the
set
of
all
manufacturing
activities
(and
mindsets)
that
are
made
possible
by
digital
technologies
– enables
people
to
do
things
they
could
not
do
before
• “sculpting”
with
3D
printer
– gives
more
opportunities
to
do
things
together
by
freely
sharing
designs
and
know-‐how
• DiDIY
Project
investigated
impact
on
several
domains
– Education
and
research
– Work
and
organisation
– Legal
systems
– Creative
Society
3. Digital
DIY
in
the
Workplace
• How
will
the
work
of
a
worker
in
a
manufacturing
firm
be
reshaped
due
to
the
influence
of
DiDIY?
– Direct
access
to
relevant
information
(e.g.
current
status
of
machines)
could
overcome
traditionally
strict
organisational
hierarchies
• Applying
an
ABM
of
an
abstract
factory
with
supervisor,
workers,
machines
and
tasks
to
investigate
the
following
research
questions:
– Allowing
workers
autonomy
in
deciding
which
task
to
do
next,
does
this
improve
the
effectiveness
of
the
production
process?
– Do
supervisors
become
superfluous?
4. StringWorld
Problem
Space
• A
world
of
strings
(things)
made
up
of
a
sequence
of
letters
(elements),
e.g.
“A”,
“AA”,
“ABAC”,
“BBC”
• Agents
(makers)
try
to
make
things
out
of
things
they
find
in
their
environment
(resources)
plus
things
they
might
get
from
other
agents,
using
a
limited
set
of
operations,
possibly
applying
specific
tools.
• Agents
aim
to
produce
certain
things
(targets)
–
using
trial
and
error
– following
a
plan
• Problem
space
to
address
DiDIY
simulation
issues
– Complex
enough
à
plans
are
worth
sharing
– still
computationally
feasible
• Prototype
implementation:
Model
of
Making
5. Factory
Model
Overview
• Workers
(agents)
realised
as
patches,
coloured
brown
• Supervisor
marked
by
red
square
• Other
patches
hold
• Resources
• Targets
• Machines
• Machines
are
tools
which
provide
a
particular
string
operation
• “add-‐B”
• “join”
• “envelope”
using
up
input
things
to
produce
a
new
output
thing
• Factory
has
to
produce
a
certain
number
of
targets
6. Specific
problems
to
solve
• Need
to
ensure
that
it
is
always
possible
to
produce
the
targets
from
the
resources
with
the
available
machine
operations
• Solution:
‘Possible
products’
network
inspired
by
firm
skills
universe
(Taylor
&
Morone
2005)
– build
network
of
nodes
(products)
and
links
(necessary
inputs),
starting
from
resources
– total
number
of
nodes
defined
by
model
parameters
num-‐resources,
num-‐targets,
num-‐
machine-‐types
7. Possible
Products
Network
Example:
3
resources,
5
machine
types,
3
targets
• Max
18
nodes
• Each
with
1,
2,
or
3
inputs
• Random
distribution
based
on
pre-‐defined
string
operations
Pick
3
of
the
5
potential
targets,
covering
all
resources
• 13,
15
and
17
Assign
operations
to
(bundles
of)
input
links
Assign
strings
to
resources
à
Derive
target
strings
8. Model
Variants
• With
supervisor
– Supervisor
assigns
jobs
to
workers
based
on
• which
target
is
the
most
outstanding
• which
machines
are
free
(for
starting
on
the
job)
– Workers
follow
the
plan
to
make
the
target
– Once
finished,
they
ask
the
supervisor
for
their
next
job
• Without
supervisor
– Workers
know
the
current
status
of
all
machines
and
which
tasks
produce
what
from
which
inputs
– Workers
decide
on
the
next
task
based
on
• which
machines
are
free
• what
things
they
have
(prefer
to
use
own
stuff
over
resources)
• Pick
most
outstanding
target
if
nothing
else
possible
10. Discussion
• First
experiments
focused
on
overall
effectiveness
of
production
process
while
varying
number
of
agents
– Production
time
(total
simulation
time
until
all
targets
achieved)
– Average
time
workers
spent
waiting
for
a
free
machine
• Introducing
simple
form
of
cooperation
– When
deliberating
possible
next
tasks,
a
worker
may
consider
not
only
the
things
(s)he
has
themselves
but
also
things
other
agents
have
• Subsequent
experiments
showed
that
results
are
very
dependent
on
the
factory
setup
(number
of
resources,
targets,
machine
types,
machines
per
type,
processing
times,
number
of
agents)
11. Introducing
Garbage
Can
Measures
• Wanted:
output
measures
to
gauge
impact
of
organisational
change
• Garbage
Can
Model:
influential
model
of
organisational
behaviour
– Problems,
participants,
opportunities,
solutions
• Three
indicators
– Problem
latency:
time
spent
by
problems
in
the
system
before
a
participant
attempts
to
solve
them
– Unsolved
problems:
number
of
problems
left
at
the
end
of
the
simulation
– Waiting
time:
time
opportunities
stay
in
the
system
waiting
to
be
used
12. GC
Indicators
for
Factory
• Translate
GC
terminology
to
factory
world
– Job
latency:
Time
spent
by
jobs
(‘problems’)
in
the
factory
before
a
worker
starts
working
on
them
– Unfinished
jobs:
number
of
unfinished
jobs
at
the
end
of
the
simulation
– Waiting
time:
Time
spent
by
free
machines
/
free
workers
waiting
to
be
used
/
start
working
on
a
new
job
13. Model
AdaptaIons
• GC
Model
assumes
streams
of
objects
• Factory
Model
– Incoming
stream
of
jobs,
with
mean
arrival
time
and
different
probabilities
for
the
different
types
– Each
job
specifies
which
target
to
produce
(a
sequence
of
tasks)
– For
cooperation,
jobs
are
split
into
the
separate
tasks
14. Preliminary
Results
(1)
Factory
1,
averages
of
20
runs
with
the
same
random
seeds,
mean
arrival
time
0.2,
simulation
stops
20
ticks
after
all
jobs
arrived
15. Preliminary
Results
(2)
Factory
1,
averages
of
20
runs
with
the
same
random
seeds,
mean
arrival
time
0.3,
simulation
stops
20
ticks
after
all
jobs
arrived
16. Preliminary
Results
(3)
Different
factory
setup
(more
complex
tasks)
with
higher
cooperation,
mean
arrival
time
0.2,
simulation
stops
20
ticks
after
all
jobs
arrived
17. Conclusion
and
Outlook
• Introduction
of
Garbage
Can
measures
helpful
in
assessing
the
factory
model
• Cooperation
manages
to
outperform
supervision
when
– Jobs
are
fairly
complex
(e.g.
with
intermediate
products
used
in
several
tasks)
– Frequency
of
jobs
is
high
• Investigation
will
continue…