Darwinbots Forum

Code center => Suggestions => Topic started by: Numsgil on March 12, 2005, 04:22:09 AM

Title: More than single cycle genes
Post by: Numsgil on March 12, 2005, 04:22:09 AM

Trying to make MB's you realize how hard it is to coordinate actions that take more than one cycle to finish.

So a waitcycle command would be cool to have.  Then we could have something like:

cond
*.robage 0 =
start
628 .aimdx store
1 waitcycle
.tie inc
1 waitcycle
.deltie inc
stop

In addition, we'd need a *.active sysvar that works similar to *.thisgene in that it changes after every gene.  *.active would store 1 if the gene is active, 0 otherwise, so you could do something like:

cond
*.active 1 !=
*.nrg 20000 >
start
20 .repro store
50 waitcycle
stop

to limit your bot from reproducing more than once every 50 cycles.

This works much smoother with the simultaneous DNA execution I was talking about in the other thread.

Title: More than single cycle genes
Post by: Zelos on March 12, 2005, 05:06:27 AM

that would help alot, make it make it, damn it make it ;)

Title: More than single cycle genes
Post by: Endy on March 13, 2005, 12:00:49 AM

Could be done with current vers but it would be massivly energy intensive.

Endy B)

Title: More than single cycle genes
Post by: PurpleYouko on March 14, 2005, 08:53:51 AM

Its not too bad to do it using timers.

The parent just needs to program the young with the same timer value that she has already. That way the whole MB is synchronized.

I'm not sure about the waitcycle gene. Not being a biologist I'm not sure how this corresponds to real life but it seems a bit iffy to me.

 :D  PY  :D

Title: More than single cycle genes
Post by: Numsgil on March 14, 2005, 09:33:02 AM

The thinking behind this is that real genes produce proteins that interact in complex ways.  These proteins hang around for longer than a single time period.

If you combine this with synchronous DNA and the timer that I mentioned earlier, you get some very powerful MB building techniques.  And perhaps some interesting SB building techniques too.

Actually, this comes from computer science more than biology.  These three together allow for multi-threaded programming.  The power of multithreading is simply incredible.

Title: More than single cycle genes
Post by: PurpleYouko on March 14, 2005, 10:25:09 AM

I don't doubt that it will work very well in DB and your example does show a similarity to real life.

Seems to me that we have to come to a decision as to whether DBs are ever going to actually simulate real life in the organic sense or if we want them to actually be little programmable robots that simulate some other fictional (or at least unknown) form of life.

They seem rather similar to replicators in StarGate.
(http://www.gateworld.net/sg1/s4/graphics/401_03.jpg)

 :D  PY  :D

Title: More than single cycle genes
Post by: Numsgil on March 14, 2005, 11:06:10 AM

Just as long as they don't try destroy the known universe, I'll be happy.  Then again, the replicators were toys at first...

Title: More than single cycle genes
Post by: PurpleYouko on March 14, 2005, 11:55:09 AM

YEEK!

Maybe they might try to invade cyberspace!

 :sly:  PY  :sly:

Title: More than single cycle genes
Post by: Old Henk on March 15, 2005, 12:13:07 PM

I think they don't need to 'simulate' life, or be our approximation (cool word, eh? B) ) of it.
They're just what they are: Darwinbots! Artificial Life! Yay!  :clap:

Title: More than single cycle genes
Post by: Endy on March 15, 2005, 04:31:09 PM

I always worry that I'll come back sometime and have them respond to me.

Endy :D

Title: More than single cycle genes
Post by: Botsareus on March 15, 2005, 08:58:29 PM

:) I still did not get smexe working the way I want, so the syberspace is safe for now :)