Author Topic: idea: use multibot at the start of a new simulation.  (Read 501 times)

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
    • DJ Paul Kononov
idea: use multibot at the start of a new simulation.
« on: July 26, 2016, 11:48:08 AM »
Mainly for evolution simulations. I had plenty of cases where I extract just the DNA and it refuses to behave correctly in a new simulation. There will be a need to rescale energy and body so the grouping of robots will have the same energy and body as a group as a single robot does at startup. Probably throw in some limits as well if it is a huge multibot it does not die from starwation: Maybe 15 body minimum and 100 nrg minimum. Also the robot that actually was identified as best will have a minimum 100 body 300 nrg.

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
    • DJ Paul Kononov
Re: idea: use multibot at the start of a new simulation.
« Reply #1 on: July 26, 2016, 05:34:29 PM »
The quickest way to recreate what I am talking about, is load a jelly fish with 'save epigenetic memory as temporary gene' enabled in global settings. Then use one of the tentacles and save the dna out. When you load the robot back in it will be a motionless tentacle.

Thinking about it further here is what I came up with. Lets say I capture a sea snake with 10 bots. That means it will be loaded with 1/10x body and energy in proportion to a sea snake loaded from scratch. What I am concerned about is maybe such a configuration is too skinny and the sea snake from scratch will win every time. Maybe instead of limits I just multiply the body and energy by log(robots)/log(2) This way the multibot version has some kind of advantage. But maybe too much of an advantage, so I play with values greater than 2 and limit it so it never goes below 1. At any rate this reduces the problem to dialing in a single value. The reason I want log in there is to stop people from cooking up multibots that have 100s of bots in them as I want to eventually expose this thing to the leagues as well.

edit: The reason I want epigenetic memory saved in the first place is one of Billys robots set its epigenetic memory and than deleted the gene that set it making extracting a functional DNA imposable.
« Last Edit: July 26, 2016, 05:53:14 PM by Botsareus »

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
    • DJ Paul Kononov
Re: idea: use multibot at the start of a new simulation.
« Reply #2 on: July 27, 2016, 10:38:16 AM »
Thinking about it even further some variables will need to be reset to not give a robot unfair advantages or disadvantages at startup. Age costs will have to be from a separate variable because some multi use 'genetic' age to build themselves, yet I do not want them to die from large age costs on startup.

Anyway, this kinds of stuff will be added only if I figure out how to fix chloroplasts. Currently it is too easy to evolve chloroplasts and too hard for a robot that only uses chloroplasts to evolve a different strategy. I have a some what working fix to make it the other way around. Unfortunately currently I am having trouble creating plants in automatic zero experiments with this 'fix.'

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
    • DJ Paul Kononov
Re: idea: use multibot at the start of a new simulation.
« Reply #3 on: July 28, 2016, 04:21:24 PM »
Off topic: If I ever get this "my version of Darwinbots" stuff going and I see a user being counterproductive I will not mess around and ban them immediately. However due to the fact that there is sooo much interest in Darwinbots these days, I do not think it is ever going to happen. :)

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
    • DJ Paul Kononov
Re: idea: use multibot at the start of a new simulation.
« Reply #4 on: October 30, 2016, 06:50:48 PM »
Yep, I actually figured out a way to solve my little problem here but it changes the way viruses work and requires a major revision. So I guess no internet mode under DB2.