Darwinbots Forum

Code center => Suggestions => Specialization, Metabolism, Digestions and Env Grid => Topic started by: Botsareus on March 21, 2005, 12:53:30 PM

Title: Lets figure it out already
Post by: Botsareus on March 21, 2005, 12:53:30 PM
How many slots of Enzymes Should we have? Num wants between 7-13 ,Shvartz wants ?70?,, How mutch you want? and why?. We will figure this out here so I can finaly get the right number into the "final virsion thread"
Title: Lets figure it out already
Post by: shvarz on March 21, 2005, 01:19:44 PM
No, it's too early for that.  There is no point to say 7 now, if in a couple of weeks we decide to have 8.  Anyway, I am thinking about 3-4 slots for "enzyme complexes" with each complex being able to hold 1-4 enzymes.  You'll see what I mean as soon as me and zelos flesh out a balanced metabolism system.
Title: Lets figure it out already
Post by: Botsareus on March 21, 2005, 01:26:56 PM
Ok , thats why I started this thread , good luck!
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 04:00:21 AM
I said this in another thread, and I thought I'd post it here too.

There should be no limit to the number of enzymes allowed.  In addition, there should be no limit to the length of these enzymes.

So why won't bots evolve ever more enzymes?  We have to charge it the more enzymes it produces.  A very non-steep exponential growth.
Title: Lets figure it out already
Post by: PurpleYouko on March 22, 2005, 10:00:03 AM
I second that. no artificaially imposed limits. just a nice big bit pattern that is able to contain a whole bunch of enzymes that may or may not overlap.

 :D  PY  :D
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 10:59:19 AM
A simple way to encourage less and shorter enzymes is to add up the length of all enzymes when deciding cost.  So there's no difference between 10 enzymes 3 bits long and one enzyme 30 bits long.

The only problem with htis, though, is that it more or less just encourages bots to have 1 enzyme that does it all, since bit patterns are more likely to be found in 1 30 bit pattern than 10 three bit patterns.
Title: Lets figure it out already
Post by: shvarz on March 22, 2005, 11:25:36 AM
Nums, here is a way to encourage splitting enzymes into different sequences:  whenever bot needs one enzyme from a string, it is charged for the whole string.
Title: Lets figure it out already
Post by: PurpleYouko on March 22, 2005, 01:36:30 PM
That works

 :D  PY  :D
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 01:42:22 PM
I would like to clarify that each bit string is one enzyme.  So yes schvarz, I agree with you.  If you need one activation site on an enzyme the whole things is produced.

But the activation sites are what do the work.  The enzymes are just the carrying mechanism for them.

Semantics really.
Title: Lets figure it out already
Post by: shvarz on March 22, 2005, 02:51:06 PM
I would call then enzyme complexes.  Yes, it's just terminology.  But mine is sooo much cooler! :)
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 02:56:50 PM
Quote
I would call then enzyme complexes.  Yes, it's just terminology.  But mine is sooo much cooler! :)
You wish! :P

The difference is that a complex implies more than one independant structures.  An enzyme complex would be a group of associated enzymes, IMO, whereas an enzyme is the single atomic structure that is produced by enzyme transcribers.
Title: Lets figure it out already
Post by: shvarz on March 22, 2005, 03:05:55 PM
Here is my proposal:  Enzymes in a single string of bits are automatically assigned to one complex.  If you get all enzymes that work on digesting fats and put them in one string, then all of them are in one complex and the whole complex can work much better than all of these enzymes on their own.

If you assemble random enzymes into a complex, then that will do you no good :)

P.S: I promise I go to work after submitting this post.
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 03:18:34 PM
I imagine the opposite as true.  An enzyme (complex to you) works by going through materials supplied to it by osmosis.  If you have alot of activation sites (enzymes to you) that do the same substance, each one will have to share the raw materials supplied to it by osmosis.

If each activation site is its own enzyme though, they can spread out and osmosis will comparitvely supply them with raw materials faster.
Title: Lets figure it out already
Post by: PurpleYouko on March 22, 2005, 04:52:26 PM
I think I have been viewing things more in line with Shvarz here.

I have been (but am not fixated on the idea) looking at enzymes as being shortish strings with a single function, contained in a single longer string (bit pattern).

I am not particularly worried about modelling the way that raw materials arrive to be sorted out. That is relatively simple to me.

Stuff arrives by whatever method (osmosis from e-grid, gathering from e-grid, eating) -----> stuff reacts with as many enzymes as are present (defining efficiency) -----> proteins, fats or whatever comes out the other end.

Shvarz's make and break system (elegant and simple) can be added as an optional extra to further increase efficiency and complexity, otherwise just let the stuff get on with it on its own with efficiency being defined purely by the number of a specific enzyme present in the bit pattern (enzyme array).

 :D  PY  :D
Title: Lets figure it out already
Post by: PurpleYouko on March 22, 2005, 05:00:30 PM
Quote
If each activation site is its own enzyme though, they can spread out and osmosis will comparitvely supply them with raw materials faster.

Don't really follow this.

Osmosis is driven by a concentration gradient across a semi-permiable membrane.
Basically the robot's body fluids will tend to equalize concentrations with the surroundings (e-grid) so to all intents and purposes the robot will contain the same concentrations as the e-grid with a slight lag. Of course this mainly applies to ionic species like salts and not to gasses like Oxygen or CO2 to such a degree.

The concentration of any of these things will be in the order of parts per million which is pretty high really.

It really shouldn't make a darned bit of difference whether the enzymes are very small seperate bit patterns or a long chain containing many activation sites. Each site will still have access to way more stuff than it can keep up with if it relies on osmosis to supply it. The stuff will be pumping in from the e-grid at a rate equal to the rate that it is being converted and will be present at a ppm level in the robot. That means several million bits of O2 will always be present and several million bits of CO2 will be present. Your enzymes will be absolutely saturated in the stuff.

 :D  PY  :D
Title: Lets figure it out already
Post by: shvarz on March 22, 2005, 05:11:07 PM
Nums, you got mme all wrong:  The thing about complex is like this:

You have a reaction chain:

A>B>C>D

You have separate enzymes for each reaction A>B, B>C, and C>D.  What happens is that each enzyme does its job and throws the product out for other enzymes to find and do their job.

Now imagine if all three enzymes are sitting in a complex and when the first enzyme is done, it hands down the product directly to second enzyme.  Second enzyme converts it to C and gives it to third enzyme.  The substrate never leaves the complex until all steps are completed.  

This is the way real organisms optimize their chemical reactions.  Sometimes the substratre is covalently linked to the complex itself until the job is done!

That's what I am calling an enzyme complex.

Now back to your bits system:

You have a string of bits:

10010001111001000011101111001100

Say enzyme 1 is coded by: 10010001
Enzyme 2: 011110010
Enzyme 3: 00111011110

Since all three enzymes are in one string, they are assigned to a single complex (just a random "flag" that is common to all of them).  Then when the reaction happens, the program will check if in the reaction A>B>C>D any enzymes are in the same complex.  It will find that they are and it will increase the efficiency of the reaction.

Clear?
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 06:00:02 PM
Musch clearer.  I agree with that.

If one complex contains more than one enzyme that does the same thing, then the substances are randomly assigned to an enzyme.  Can we agree with that too?
Title: Lets figure it out already
Post by: Botsareus on March 22, 2005, 06:05:22 PM
Ok, so how many slots enzymes we are going to have? How is it determined?
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 06:10:03 PM
Infinitely many, as I've already said atleast twice.
Title: Lets figure it out already
Post by: Botsareus on March 22, 2005, 06:11:07 PM
How is it determined?
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 06:14:20 PM
Pronoun trouble.  What is the 'it' to which you refer?
Title: Lets figure it out already
Post by: Botsareus on March 22, 2005, 06:17:11 PM
lol sry Num, In Programming there is no such thing as an “Infinite” variable; How does the robot figure out how many enzyme slots it needs?
I mean it does not start out having “Infinite” enzyme slots right?
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 06:20:00 PM
When you say enzyme slot, what exactly do you mean?  As I understand it now the question doesn't make alot of sense.
Title: Lets figure it out already
Post by: Botsareus on March 22, 2005, 06:30:20 PM
http://s9.invisionfree.com/DarwinBots_Foru...topic=187&st=15 (http://s9.invisionfree.com/DarwinBots_Forum/index.php?showtopic=187&st=15)


Numsgil:
Quote
Okay, so we're basically in agreement to limit the total number of enzymes a bot can have? I'm not sure how enzyme length and cost should correlate. Either linearly or exponentially, or maybe some other system.

20 may be too many enzyme slots. I'm really not sure. I really have no idea how natural selection would affect them. Having a single enzyme that can be infinitely long is an interesting idea. I would say something like 7-13 enzyme slots would be about right, depending on how in depth and detailed we make digestion.

Funny, now Numsgil tells me he does not know what an enzyme slot is....
Title: Lets figure it out already
Post by: Numsgil on March 22, 2005, 06:35:41 PM
Oh, okay.  That's what I thought.  Infinite.  That is, as many slots as you want.  There's no limit.  You can have 3, 500, or a million.  (Well, a million is overkill).
Title: Lets figure it out already
Post by: Botsareus on March 22, 2005, 06:37:00 PM
ok , ok I get it , thx