Evolution algorithm development

handy · December 22, 2024, 3:57pm

I am new to C# and game engines but I have some experience in python and quite the experience in evolution algorithm, so I attempted to understand the current code source but got Lost , so I want some guidance towards the right place to find the current code of EA(evolution algorithm)

aah31415 · December 22, 2024, 4:09pm

The evolution algorithm of Thrive is called “autoevo” or “auto-evo” if that helps.

C009 · December 22, 2024, 4:24pm

I can point you to line 703 of GameWorld, this is one place where auto-evo get’s called. Thrive/src/general/base_stage/EditorBase.cs at master · Revolutionary-Games/Thrive · GitHub
I recommend exploring stuff with some IDE, because usually you can right click method call and you can then go to method implementation or all method calls.

aah31415 · December 22, 2024, 4:39pm

What does “IDE” stand for?

C009 · December 22, 2024, 5:25pm

Integrated Development Environment.

hhyyrylainen · December 22, 2024, 5:35pm

@handy As a quick summary of all core concepts of auto-evo, this document is a must read:

github.com

Revolutionary-Games/Thrive/blob/master/doc/auto_evo.md

# Auto-Evo

Auto-Evo is not a single algorithm, but a longstanding problem at the center of
thrive. The goal of Auto-Evo is for Thrive to be able to generate new creatures
to populate the world through the evolutionary process in real time. This
problem comes with a few constraints. As already mentioned, Auto-Evo must be
able to run in real time on all of our player's PCs. Additionally, Auto-Evo must
generate creatures that have sufficient variety to provide an interesting
experience for the player, while still sticking to expectations of a viable
creature.

## The Miche Algorithm
While Auto-Evo is not in itself an algorithm, it is solved by one. The current
algorithm used to solve Auto-Evo is the Miche Algorithm.

### Miche
The term Miche comes from a shortening of Multi-Niche (and a nod to computer
science professor Dr. Meesh.) Each Miche is a node in a tree structure called
the Miche Tree. Each of these Miches contains a list of child Miches,
recursively creating the tree. If a node does not contain any child nodes it is

This file has been truncated. show original

While it’s possible to navigate from here to the code you need to see. I think a better starting point is looking specifically at the files in the auto-evo folder:

And while just randomly checking things starting from the top level will eventually work, I would also recommend using an IDE and starting here from the population number calculation part and jumping to the called methods to start to get into the parts of the auto-evo system:

github.com

Revolutionary-Games/Thrive/blob/master/src/auto-evo/simulation/MichePopulation.cs#L191


      
                  // Simulate the species in each patch taking into account the already computed populations
                  SimulatePatchStep(parameters, entry.Value,
                      species.Where(s => parameters.Results.GetPopulationInPatch(s, entry.Value) > 0), insertWorkingMemory,
                      random, cache);
              }
          }
          
          /// <summary>
          ///   The heart of the simulation that handles the processed parameters and calculates future populations.
          /// </summary>
          private static void SimulatePatchStep(SimulationConfiguration simulationConfiguration, Patch patch,
              IEnumerable<Species> genericSpecies, Miche.InsertWorkingMemory insertWorkingMemory, Random random,
              SimulationCache cache)
          {
              _ = random;
          
              var populations = simulationConfiguration.Results;
              bool trackEnergy = simulationConfiguration.CollectEnergyInformation;
          
              // Note that this modifies the miche tree while simulating
              var miche = populations.GetMicheForPatch(patch);

The other key part is then the miche tree generation and the selection pressures that are on the miche tree nodes.

aah31415 · December 29, 2024, 1:24pm

So, handy, have you managed to understand the auto-evo algorithm yet? You didn’t ask us for any more clues, so I’d guess you didn’t stumble upon any more major problems…

handy · December 29, 2024, 1:49pm

Rather then not stumbling into any major problem , I would rather define my current stat as stumbling into thrive , I am at a critical year in my education (I have an international Exam at the end of the year), so I am obliged to focus on my studies, never the less, I am trying to read the documents mentioned in the doc repertory, I might as well try to understand game engines in the summer holidays, moreover I am a perfectionist guy so I might as well understand everything there is to understand about the game before attempting my first contribution, anyway since I am quite busy{studying:( }, I would be very happy if you always send me notification about the current discussion and issues of the thrive community. Because I really want to participate in some of the theoratical discussion about auto-evo and the realistic aspect of the game

aah31415 · December 29, 2024, 2:08pm

You should focus on your Life before Thrive. If you can’t find time to contribute something in, there surely should be somebody else to do that. It’s just that it makes Thrive’s development take longer…

handy · December 29, 2024, 2:27pm

Well taking long time is not a problem,I mean as long as the game is attracting more and more attention and as long as this attention is converted into a contribution things will go smoothly, that’s why I am into long term contribution, if I start learning game engines then Godot specifically, then honing my skills in C#, when I aquire the needed skills it will be easier to contribute actively rather than just talking and day dreaming

aah31415 · December 29, 2024, 4:14pm

Agreed. That is the type of developers we need.