16 replies to this topic

[BitTamer]

If anyone has ever read a book called Prey by Michael Crichton,

Is there any specific ai algorithm for "distributed processing" like the swarms in the book?
Has anyone ever done it?

The swarms should be able to evolve just like in the book.

If you haven't read the book, then basically the swarms are nanoparticles which are programmed to simulate predators. Only fundamental rules are programmed into each particle and as a whole, the outcome is unpredictable.

[dave_]

sounds like Cellular automata/emergent behaviour, but what's your application?

[Rubicon]

I've read it, and boy did it suck! He peaked with Andromeda Strain imo :)

The tenet is that the thingies were meant to conglomerate around each other and form a camera, but it's a book and it went wrong and started eating people.

The idea to control the particles using simple rules to generate emergent behaviour, but somehow controlling that emergence to give a predictably useful behaviour (like focussing a fly's eye camera) is very interesting.

In practice, I don't think nano will ever take off, much as I'd love to see it happen. Any sort of limited cpu processing, sensors and moving parts at all would make them so large as to be unuseable in most fields romanticists envisage. I personally can't see any "nano" bots ever getting smaller than a pinhead. Freakin small, but I don't want em in my brain! :)

[BitTamer]

What I dont get is, since the nanoparticles are so small, how can they be programmed in the first place?

[Skavenger]

via very small cables.

[dave_]

What I dont get is, since transistors in CPUs are so small how can they be programmed in the first place?

There's clearly a lot you don't get. Why not read some articles about nano-robots and nano-technology? This isnt a nano tech website, its a games programming one.

[BitTamer]

Cables?
You mean very small computer chips embedded into each particle?
But they're way too small, aren't they?

[dave_]

BitTamer said:

Cables?
You mean very small computer chips embedded into each particle?
But they're way too small, aren't they?

An average human hair is 80000nm wide, a red blood cell is ~7000nm
An average feature on a Intel core duo is 65nm.

[BitTamer]

So if i use swarm AI in my application, would it slow it down?
Cos each particle would carry out its own processes.

[dave_]

Would what slow down? Slow down compared to what? Such a vague question.

Swarm AI is fiction in this context.
You need to learn how to do research. A good starting point for research is google. First hit is this, for Swarm AI. It even refers to your book.
Here is an example of its application: boids

[Rubicon]

>>An average feature on a Intel core duo is 65nm.

And a hundred of those maketh not a machine

[Rubicon]

>> You need to learn how to do research. A good starting point for research is google.

Come on, play fair. He's obviously a n00b, but a good way to research stuff is also to ask those who know something about the subject some starting point questions.

I've been researching smooth particle hydrodynamics a lot recently and started with google. Fifty stuffy academic papers full of foot long equations later, I asked in gdalgorithms and was put straight instantly.

[dave_]

Rubicon said:

>>An average feature on a Intel core duo is 65nm.
And a hundred of those maketh not a machine

No I never suggested that, and anyway you're several order of magnitude out. The figures are diameters.

You'd be able to fit 1,500,000 things of 65nm diameter in the area of a cross-section of human hair.
Even assuming a transistor size of 6λ thats approximately 42,000 transistors, which is more than enough to create something simple, a z80 had a mere 6000 transistors.

You'd still be able to fit hundreds in the size of a red blood cell.

And the idea is not to create a general purpose CPU, but a simple machine with few rules that acts in unison with other particles.

But anyway, like I said and you clearly do: Google/wikipedia first ask then ask questions when you don't understand something.

[Rubicon]

yawn

[dave_]

Is being wrong tiring?

[Rubicon]

Nope, that's arguing with egomaniacs on message boards who just go on and on and on and must always have the last say.

Cue you to say something witty like "you mean like you just did?".....

[BitTamer]

Alright, lets stop the arguing.
I'll read a little about it then, shall I?
I just wanted to talk to someone who knows something about it cos i find the tutorials on the net too long and unclear.