|
|
| Supervisor: |
DR.
DANIEL COORE |
| Students: |
Ricardo
Anderson, Jamala Bryan, John Muirhead, Richard Lawson, Howard Nation,
Vernon Rowe, Daryl Strachan |
| |
|
|
|
OVERVIEW
|
|
WORK
IN PROGRESS
|
| A
colony of cells, sharing a common genetic code, self-organises to
produce an organism. Amorphous Computing seeks to understand the
organisational principles behind such phenomena by studying methods for
programming computational models of such systems.
|
|
GPL:
A language for specifying patterns. Programs are written in terms of
"growing points" that can move across the substrate of particles,
depositing material which we view as a pattern. |
| |
 |
| |
 |
The
evolution of a GPL program producing the pattern at the far left. This
pattern represents a CMOS layout of an inverter, one of the most
primitive circuit elements of digital logic. |
| |
AIM
To
be able to configure (program) complex systems of locally interacting
elements to achieve a prespecified emergent
behaviour. |
|
|
|
|
|
|
| |
|
|
|
|
|
| |
ECOLI:
A language for specifying interactions. Programs are written in terms
of responses to events (messages and inputs). A response on one
particle triggers an event on its neighbours, and they respond in turn.
Programs at this level try to control explosion events. |
| |
| |
 |
 |
| MODEL |
|
| Particles
are irregularly located in space. They do not know their own
coordinates. |
 |
|
| |
 |
Each
processor runs the same program. In this program, one processor
triggers its neighbor to send a value in its message that is one higher
than that received. The intended result is that processors label
themselves with the hop-count distance from the initiating processor. |
|
The
evolution of the program above. Source points randomly select
themselves to initiate the first set of messages. Each colour in the
images represents a different distance. |
| |
| |
| |
| |
|
|
|
|
| Particles
communicate by broadcasting over a fixed range.
|
 |
|
|
|
| |
Simulations
of
physical systems enable us to explore our ideas realistically.
Amorphous Computers might be realised in many ways, e.g: silicon based
chips, bacteria, nano-machines, or software agents acting in a virtual
system. The low level details of each of these systems requires a
different implementation. We strive to design flexible and powerful
Amorphous Computing simulators without sacrificing performance. |
| |
| |
| Particles
have small memories and simple processing power. |
|
 |
 |
| |
| Particles
receive input by sensing their environment. |
|
| |
| All
particles run the same program. Behavioural differentiation arises from
sensor changes and inter-particle communication. |
|
| |
| |
|
A
diagram of a proposed protocol for inter-particle communication. This
protocol can be embedded into the ECOLI-interpreting simulator.
|
| |
|
| Particles
run asynchronously, but with comparable speeds. |
|
|
|
A
screen shot of ECOLI-interpreting simulator, developed in Summer 2001 |
| |
|
|
| |
|
|
|
|
| POTENTIAL
APPLICATIONS |
|
|
|
|
|
| "Smart"
Structures, e.g. |
|
|
Manipulating
micro-organisms |
|
Providing
Models for: |
- stronger
bridges
- active
aeroplane wings
- ergonomically
sensitive furniture
- roads
that report traffic loads
|
|
- make
molecular-scale electronic circuits
- tag
diseased cells
- dispense
drugs to localised sites
- provide
data storage
|
|
-
low
cost supercomputing
- large-scale
resource management
|
| PROJECTS |
|
|
|
|
