Added: 2 years ago
From: AntiRushx2
Views: 10,146
Sort by time | Sort by thread (beta)

Link to this comment:

Share to:

All Comments (22)

Sign In or Sign Up now to post a comment!
  • Very cool. what did your chromosomes contain? the wall-hugging behavior was interesting

  • yeah I count 9 vs 6 units, unfair...

    more numbers of same unit...

  • @PlebScrubber Yeah, but can you fight with the same number of units and still end up with 5 like he did? Maybe... Let's tell him to keep evolving, but to mix this genetic code with another 5 which each have 500 generations then to mix all of them together and make a smart one.

  • @TimJSwan89 all it does it have many units attack the same enemy unit, so it dies very quickly, and then all target the next one, this means the total damage taken is much less and more units survive

    if you left the units to act naturally, they would each just target the nearest enemy and their combined fire would be spread thin, that means enemy units will take longer to die, and have more time to deal damage to friendly troops

  • @PlebScrubber count again...

  • 9 vs 6? Unfair...

  • @UnrealQW

    There were nine units on each side, but red won because three of blue's units showed up late to the fight.

  • @tifforo1

    blue also does several actions to micro the dieing units away making red target the other higher hp unit

  • @tifforo1

    red also does several actions to micro the dieing units away making blue target the other higher hp unit

  • so this was an evolutionary arms race with red vs blue? or was red evolving the best solution to attack blue?

  • Are there any problems that can't be solved with genetic algorithms?

  • @ross817 Finding a girlfriend

  • @hmanham

    this problem was solved by many products of genetic algorithms

  • @ross817 Yes, most problems can't be solved with genetic algorithms. However, if you have a way to judge how good a solution is and when the problem is solved, and the factors aren't too many, it should be solveable with genetic algorithms.

  • is this comp ai

    or is this like... you program ur own units to do better in matches? xD

  • This is very impressive. Came from your site.

  • Very cool! How many generations did it take to come up with this solution?

  • @MiniAgupa This was generation 638. Similar solutions started to emerge after ~400.

  • @AntiRushx2 Can you provide your compiled AI ? I wanna play vs it.

    I have a cpu than run day and night if you want to let it learn an infinite number of time

    @FusionNinjin in c++, using bwapi

  • nifty, but i think a better question is how you can program untop of starcraft :S

  • how does your algorithm work ?

  • @igoronline

    Like the title says: Genetic Programming. The agents genetically inherit the traits that maximize their fitness (survival).

Loading...
Alert icon
0 / 00Unsaved Playlist Return to active list
    1. Your queue is empty. Add videos to your queue using this button:
      or sign in to load a different list.
    Loading...Loading...Saving...
    • Clear all videos from this list
    • Learn more