search

mulander / evolutionchamber

Automatically exported from code.google.com/p/evolutionchamber
0 stars 0 forks source link

2 Spawning Pools #85

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Some builds are coming back satisfied with 2 pools.  There is no reason for a 
player to ever build 2 pools.

The second pool came after the WP2 (7~ish time) in this build:

My settings:

Final:
8 processor
time limit 120
1 overseer
3 queen
10 zerglings
metabolic boost
6 mutalisks
1 lair
2 extractors
1 pool
1 spire

WP1:
time 4:30
1 queen
6 zerglings
metabolic boost

WP2:
time 7:00
1 overseer
1 lair

Original issue reported on code.google.com by phezm...@gmail.com on 26 Oct 2010 at 5:32

GoogleCodeExporter commented 9 years ago 
yes, even if there are 2 upgrades i do not see any build that will research 
those both in the same time. SP should be limited to 1.
The same for other buildings if you do not select multiple upgrades.
Another common thing i sow is 3 evochambers when i select only +2 armor.
I think limiting buildings based on the input can speed things a little.

Original comment by ice...@gmail.com on 27 Oct 2010 at 7:32

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Putting 1 Spawning Pool on final also doesn't fix the issue, it makes no sense 
why it would make more than 1 Spawning Pool unless the user specifically said 2.

Original comment by sean.cow...@gmail.com on 2 Nov 2010 at 7:29

GoogleCodeExporter commented 9 years ago 
Perhaps this could be solved by adding constraints, such that in the same way 
we set a target amount, we could set a maximum. 

Providing constraints on the algorithm could help exclude poor branches before 
even attempting them.

Basically a way to say "at most 1 spawning pool <or _____ unit/building>"

Original comment by hugh.lo...@lodestarbpm.com on 2 Nov 2010 at 8:19

GoogleCodeExporter commented 9 years ago 
Issue 111 has been merged into this issue.

Original comment by azzur...@gmail.com on 3 Nov 2010 at 8:06

GoogleCodeExporter commented 9 years ago 
In genetic algorithms, there are sometimes un-intuitive jumps from good builds 
to bad builds, to an even better build.

So, while that will never absolutely go away, I believe we are currently 
ignoring more than one spawning pool in the actual build.

Original comment by Frit...@gmail.com on 6 Nov 2010 at 1:12