<s>
Swarm	B-Architecture
intelligence	I-Architecture
(	O
SI	O
)	O
is	O
the	O
collective	O
behavior	O
of	O
decentralized	O
,	O
self-organized	O
systems	O
,	O
natural	O
or	O
artificial	O
.	O
</s>
<s>
The	O
concept	O
is	O
employed	O
in	O
work	O
on	O
artificial	B-Application
intelligence	I-Application
.	I-Application
</s>
<s>
SI	O
systems	O
consist	O
typically	O
of	O
a	O
population	O
of	O
simple	O
agents	B-General_Concept
or	O
boids	B-Application
interacting	O
locally	O
with	O
one	O
another	O
and	O
with	O
their	O
environment	O
.	O
</s>
<s>
The	O
agents	B-General_Concept
follow	O
very	O
simple	O
rules	O
,	O
and	O
although	O
there	O
is	O
no	O
centralized	O
control	O
structure	O
dictating	O
how	O
individual	O
agents	B-General_Concept
should	O
behave	O
,	O
local	O
,	O
and	O
to	O
a	O
certain	O
degree	O
random	O
,	O
interactions	O
between	O
such	O
agents	B-General_Concept
lead	O
to	O
the	O
emergence	O
of	O
"	O
intelligent	O
"	O
global	O
behavior	O
,	O
unknown	O
to	O
the	O
individual	O
agents	B-General_Concept
.	O
</s>
<s>
Examples	O
of	O
swarm	B-Architecture
intelligence	I-Architecture
in	O
natural	O
systems	O
include	O
ant	B-Algorithm
colonies	O
,	O
bee	O
colonies	O
,	O
bird	O
flocking	O
,	O
hawks	O
hunting	O
,	O
animal	O
herding	O
,	O
bacterial	O
growth	O
,	O
fish	O
schooling	O
and	O
microbial	O
intelligence	O
.	O
</s>
<s>
The	O
application	O
of	O
swarm	O
principles	O
to	O
robots	B-Algorithm
is	O
called	O
swarm	O
robotics	O
while	O
swarm	B-Architecture
intelligence	I-Architecture
refers	O
to	O
the	O
more	O
general	O
set	O
of	O
algorithms	O
.	O
</s>
<s>
Boids	B-Application
is	O
an	O
artificial	O
life	O
program	O
,	O
developed	O
by	O
Craig	O
Reynolds	O
in	O
1986	O
,	O
which	O
simulates	O
flocking	O
was	O
published	O
in	O
1987	O
in	O
the	O
proceedings	O
of	O
the	O
ACM	O
SIGGRAPH	O
conference	O
.	O
</s>
<s>
As	O
with	O
most	O
artificial	O
life	O
simulations	O
,	O
Boids	B-Application
is	O
an	O
example	O
of	O
emergent	O
behavior	O
;	O
that	O
is	O
,	O
the	O
complexity	O
of	O
Boids	B-Application
arises	O
from	O
the	O
interaction	O
of	O
individual	O
agents	B-General_Concept
(	O
the	O
boids	B-Application
,	O
in	O
this	O
case	O
)	O
adhering	O
to	O
a	O
set	O
of	O
simple	O
rules	O
.	O
</s>
<s>
The	O
rules	O
applied	O
in	O
the	O
simplest	O
Boids	B-Application
world	O
are	O
as	O
follows	O
:	O
</s>
<s>
as	O
a	O
special	O
case	O
of	O
the	O
boids	B-Application
model	O
introduced	O
in	O
1986	O
by	O
Reynolds	O
.	O
</s>
<s>
Evolutionary	B-Algorithm
algorithms	I-Algorithm
(	O
EA	O
)	O
,	O
particle	B-Algorithm
swarm	I-Algorithm
optimization	I-Algorithm
(	O
PSO	O
)	O
,	O
differential	B-Algorithm
evolution	I-Algorithm
(	O
DE	O
)	O
,	O
ant	B-Algorithm
colony	I-Algorithm
optimization	I-Algorithm
(	O
ACO	O
)	O
and	O
their	O
variants	O
dominate	O
the	O
field	O
of	O
nature-inspired	O
metaheuristics	B-Algorithm
.	O
</s>
<s>
A	O
large	O
number	O
of	O
more	O
recent	O
metaphor-inspired	O
metaheuristics	B-Algorithm
have	O
started	O
to	O
attract	O
criticism	O
in	O
the	O
research	O
community	O
for	O
hiding	O
their	O
lack	O
of	O
novelty	O
behind	O
an	O
elaborate	O
metaphor	O
.	O
</s>
<s>
For	O
algorithms	O
published	O
since	O
that	O
time	O
,	O
see	O
List	B-Algorithm
of	I-Algorithm
metaphor-based	I-Algorithm
metaheuristics	I-Algorithm
.	O
</s>
<s>
Metaheuristics	B-Algorithm
lack	O
a	O
confidence	O
in	O
a	O
solution	O
.	O
</s>
<s>
One	O
such	O
instance	O
is	O
Ant	B-Algorithm
inspired	O
Monte	O
Carlo	O
algorithm	O
for	O
Minimum	O
Feedback	O
Arc	O
Set	O
where	O
this	O
has	O
been	O
achieved	O
probabilistically	O
via	O
hybridization	O
of	O
Monte	O
Carlo	O
algorithm	O
with	O
Ant	B-Algorithm
Colony	I-Algorithm
Optimization	I-Algorithm
technique	O
.	O
</s>
<s>
First	O
published	O
in	O
1989	O
Stochastic	B-Algorithm
diffusion	I-Algorithm
search	I-Algorithm
(	O
SDS	O
)	O
was	O
the	O
first	O
Swarm	B-Architecture
Intelligence	I-Architecture
metaheuristic	B-Algorithm
.	O
</s>
<s>
SDS	O
is	O
an	O
agent-based	O
probabilistic	B-General_Concept
global	O
search	O
and	O
optimization	O
technique	O
best	O
suited	O
to	O
problems	O
where	O
the	O
objective	O
function	O
can	O
be	O
decomposed	O
into	O
multiple	O
independent	O
partial-functions	O
.	O
</s>
<s>
Unlike	O
the	O
stigmergic	O
communication	O
used	O
in	O
ACO	O
,	O
in	O
SDS	O
agents	B-General_Concept
communicate	O
hypotheses	O
via	O
a	O
one-to-one	O
communication	O
strategy	O
analogous	O
to	O
the	O
tandem	O
running	O
procedure	O
observed	O
in	O
Leptothorax	O
acervorum	O
.	O
</s>
<s>
A	O
positive	O
feedback	O
mechanism	O
ensures	O
that	O
,	O
over	O
time	O
,	O
a	O
population	O
of	O
agents	B-General_Concept
stabilise	O
around	O
the	O
global-best	O
solution	O
.	O
</s>
<s>
Recent	O
work	O
has	O
involved	O
merging	O
the	O
global	O
search	O
properties	O
of	O
SDS	O
with	O
other	O
swarm	B-Architecture
intelligence	I-Architecture
algorithms	O
.	O
</s>
<s>
Ant	B-Algorithm
colony	I-Algorithm
optimization	I-Algorithm
(	O
ACO	O
)	O
,	O
introduced	O
by	O
Dorigo	O
in	O
his	O
doctoral	O
dissertation	O
,	O
is	O
a	O
class	O
of	O
optimization	O
algorithms	O
modeled	O
on	O
the	O
actions	O
of	O
an	O
ant	B-Algorithm
colony	O
.	O
</s>
<s>
ACO	O
is	O
a	O
probabilistic	B-General_Concept
technique	I-General_Concept
useful	O
in	O
problems	O
that	O
deal	O
with	O
finding	O
better	O
paths	O
through	O
graphs	O
.	O
</s>
<s>
Artificial	O
'	O
ants'—simulation	O
agents	B-General_Concept
—	O
locate	O
optimal	O
solutions	O
by	O
moving	O
through	O
a	O
parameter	O
space	O
representing	O
all	O
possible	O
solutions	O
.	O
</s>
<s>
Particle	B-Algorithm
swarm	I-Algorithm
optimization	I-Algorithm
(	O
PSO	O
)	O
is	O
a	O
global	O
optimization	O
algorithm	O
for	O
dealing	O
with	O
problems	O
in	O
which	O
a	O
best	O
solution	O
can	O
be	O
represented	O
as	O
a	O
point	O
or	O
surface	O
in	O
an	O
n-dimensional	O
space	O
.	O
</s>
<s>
The	O
main	O
advantage	O
of	O
such	O
an	O
approach	O
over	O
other	O
global	O
minimization	O
strategies	O
such	O
as	O
simulated	B-Algorithm
annealing	I-Algorithm
is	O
that	O
the	O
large	O
number	O
of	O
members	O
that	O
make	O
up	O
the	O
particle	B-Algorithm
swarm	I-Algorithm
make	O
the	O
technique	O
impressively	O
resilient	O
to	O
the	O
problem	O
of	O
local	O
minima	O
.	O
</s>
<s>
Artificial	B-Architecture
Swarm	I-Architecture
Intelligence	I-Architecture
(	O
ASI	O
)	O
is	O
method	O
of	O
amplifying	O
the	O
collective	O
intelligence	O
of	O
networked	O
human	O
groups	O
using	O
control	O
algorithms	O
modeled	O
after	O
natural	O
swarms	O
.	O
</s>
<s>
Sometimes	O
referred	O
to	O
as	O
Human	O
Swarming	O
or	O
Swarm	O
AI	B-Application
,	O
the	O
technology	O
connects	O
groups	O
of	O
human	O
participants	O
into	O
real-time	O
systems	O
that	O
deliberate	O
and	O
converge	O
on	O
solutions	O
as	O
dynamic	O
swarms	O
when	O
simultaneously	O
presented	O
with	O
a	O
question	O
ASI	O
has	O
been	O
used	O
for	O
a	O
wide	O
range	O
of	O
applications	O
,	O
from	O
enabling	O
business	O
teams	O
to	O
generate	O
highly	O
accurate	O
financial	O
forecasts	O
to	O
enabling	O
sports	O
fans	O
to	O
outperform	O
Vegas	O
betting	O
markets	O
.	O
</s>
<s>
The	O
U.S.	O
military	O
is	O
investigating	O
swarm	B-Architecture
techniques	I-Architecture
for	O
controlling	O
unmanned	O
vehicles	O
.	O
</s>
<s>
A	O
1992	O
paper	O
by	O
M	O
.	O
Anthony	O
Lewis	O
and	O
George	O
A	O
.	O
Bekey	O
discusses	O
the	O
possibility	O
of	O
using	O
swarm	B-Architecture
intelligence	I-Architecture
to	O
control	O
nanobots	O
within	O
the	O
body	O
for	O
the	O
purpose	O
of	O
killing	O
cancer	O
tumors	O
.	O
</s>
<s>
Conversely	O
al-Rifaie	O
and	O
Aber	O
have	O
used	O
stochastic	B-Algorithm
diffusion	I-Algorithm
search	I-Algorithm
to	O
help	O
locate	O
tumours	O
.	O
</s>
<s>
Swarm	B-Architecture
intelligence	I-Architecture
has	O
also	O
been	O
applied	O
for	O
data	B-Application
mining	I-Application
and	O
cluster	B-Algorithm
analysis	I-Algorithm
.	O
</s>
<s>
Ant	B-Algorithm
based	O
models	O
are	O
further	O
subject	O
of	O
modern	O
management	O
theory	O
.	O
</s>
<s>
The	O
use	O
of	O
swarm	B-Architecture
intelligence	I-Architecture
in	O
telecommunication	O
networks	O
has	O
also	O
been	O
researched	O
,	O
in	O
the	O
form	O
of	O
ant-based	B-Algorithm
routing	I-Algorithm
.	O
</s>
<s>
Basically	O
,	O
this	O
uses	O
a	O
probabilistic	B-General_Concept
routing	O
table	O
rewarding/reinforcing	O
the	O
route	O
successfully	O
traversed	O
by	O
each	O
"	O
ant	B-Algorithm
"	O
(	O
a	O
small	O
control	O
packet	O
)	O
which	O
flood	O
the	O
network	O
.	O
</s>
<s>
Mobile	O
media	O
and	O
new	O
technologies	O
have	O
the	O
potential	O
to	O
change	O
the	O
threshold	O
for	O
collective	O
action	O
due	O
to	O
swarm	B-Architecture
intelligence	I-Architecture
(	O
Rheingold	O
:	O
2002	O
,	O
P175	O
)	O
.	O
</s>
<s>
A	O
very	O
different-ant	O
inspired	O
swarm	B-Architecture
intelligence	I-Architecture
algorithm	O
,	O
stochastic	B-Algorithm
diffusion	I-Algorithm
search	I-Algorithm
(	O
SDS	O
)	O
,	O
has	O
been	O
successfully	O
used	O
to	O
provide	O
a	O
general	O
model	O
for	O
this	O
problem	O
,	O
related	O
to	O
circle	O
packing	O
and	O
set	O
covering	O
.	O
</s>
<s>
Airlines	O
have	O
also	O
used	O
ant-based	B-Algorithm
routing	I-Algorithm
in	O
assigning	O
aircraft	O
arrivals	O
to	O
airport	O
gates	O
.	O
</s>
<s>
At	O
Southwest	O
Airlines	O
a	O
software	O
program	O
uses	O
swarm	B-Architecture
theory	I-Architecture
,	O
or	O
swarm	B-Architecture
intelligence	I-Architecture
—	O
the	O
idea	O
that	O
a	O
colony	O
of	O
ants	O
works	O
better	O
than	O
one	O
alone	O
.	O
</s>
<s>
Each	O
pilot	O
acts	O
like	O
an	O
ant	B-Algorithm
searching	O
for	O
the	O
best	O
airport	O
gate	O
.	O
</s>
<s>
Stanley	O
and	O
Stella	O
in	O
:	O
Breaking	O
the	O
Ice	O
was	O
the	O
first	O
movie	O
to	O
make	O
use	O
of	O
swarm	O
technology	O
for	O
rendering	O
,	O
realistically	O
depicting	O
the	O
movements	O
of	O
groups	O
of	O
fish	O
and	O
birds	O
using	O
the	O
Boids	B-Application
system	O
.	O
</s>
<s>
Tim	O
Burton	O
's	O
Batman	B-Application
Returns	I-Application
also	O
made	O
use	O
of	O
swarm	O
technology	O
for	O
showing	O
the	O
movements	O
of	O
a	O
group	O
of	O
bats	O
.	O
</s>
<s>
Airlines	O
have	O
used	O
swarm	B-Architecture
theory	I-Architecture
to	O
simulate	O
passengers	O
boarding	O
a	O
plane	O
.	O
</s>
<s>
Southwest	O
Airlines	O
researcher	O
Douglas	O
A	O
.	O
Lawson	O
used	O
an	O
ant-based	O
computer	O
simulation	O
employing	O
only	O
six	O
interaction	O
rules	O
to	O
evaluate	O
boarding	O
times	O
using	O
various	O
boarding	O
methods	O
.	O
</s>
<s>
The	O
study	O
showed	O
a	O
23%	O
increase	O
in	O
diagnostic	O
accuracy	O
when	O
using	O
Artificial	B-Architecture
Swarm	I-Architecture
Intelligence	I-Architecture
(	O
ASI	O
)	O
technology	O
compared	O
to	O
majority	O
voting	O
.	O
</s>
<s>
These	O
grammars	O
interact	O
as	O
agents	B-General_Concept
behaving	O
according	O
to	O
rules	O
of	O
swarm	B-Architecture
intelligence	I-Architecture
.	O
</s>
<s>
Such	O
behavior	O
can	O
also	O
suggest	O
deep	B-Algorithm
learning	I-Algorithm
algorithms	O
,	O
in	O
particular	O
when	O
mapping	O
of	O
such	O
swarms	O
to	O
neural	O
circuits	O
is	O
considered	O
.	O
</s>
<s>
have	O
successfully	O
used	O
two	O
swarm	B-Architecture
intelligence	I-Architecture
algorithms	O
—	O
one	O
mimicking	O
the	O
behaviour	O
of	O
one	O
species	O
of	O
ants	O
(	O
Leptothorax	O
acervorum	O
)	O
foraging	O
(	O
stochastic	B-Algorithm
diffusion	I-Algorithm
search	I-Algorithm
,	O
SDS	O
)	O
and	O
the	O
other	O
algorithm	O
mimicking	O
the	O
behaviour	O
of	O
birds	O
flocking	O
(	O
particle	B-Algorithm
swarm	I-Algorithm
optimization	I-Algorithm
,	O
PSO	O
)	O
—	O
to	O
describe	O
a	O
novel	O
integration	O
strategy	O
exploiting	O
the	O
local	O
search	O
properties	O
of	O
the	O
PSO	O
with	O
global	O
SDS	O
behaviour	O
.	O
</s>
<s>
The	O
resulting	O
hybrid	B-Algorithm
algorithm	I-Algorithm
is	O
used	O
to	O
sketch	O
novel	O
drawings	O
of	O
an	O
input	O
image	O
,	O
exploiting	O
an	O
artistic	O
tension	O
between	O
the	O
local	O
behaviour	O
of	O
the	O
'	O
birds	O
flocking'—as	O
they	O
seek	O
to	O
follow	O
the	O
input	O
sketch	O
—	O
and	O
the	O
global	O
behaviour	O
of	O
the	O
"	O
ants	O
foraging	O
"	O
—	O
as	O
they	O
seek	O
to	O
encourage	O
the	O
flock	O
to	O
explore	O
novel	O
regions	O
of	O
the	O
canvas	O
.	O
</s>
<s>
The	O
"	O
creativity	O
"	O
of	O
this	O
hybrid	O
swarm	B-Architecture
system	I-Architecture
has	O
been	O
analysed	O
under	O
the	O
philosophical	O
light	O
of	O
the	O
"	O
rhizome	O
"	O
in	O
the	O
context	O
of	O
Deleuze	O
's	O
"	O
Orchid	O
and	O
Wasp	O
"	O
metaphor	O
.	O
</s>
<s>
freedom	O
and	O
constraints	O
)	O
within	O
the	O
swarm	B-Architecture
intelligence	I-Architecture
's	O
two	O
infamous	O
phases	O
of	O
exploration	O
and	O
exploitation	O
.	O
</s>
