Genetic AI Project

This sounds like an interesting project but I simply can't get it to work. I put GeneticAI into the MODS folder in My Documents and put an AI datafile within the GeneticAI folder. It crashes when loading the map each time.
 
Moham said:
This sounds like an interesting project but I simply can't get it to work. I put GeneticAI into the MODS folder in My Documents and put an AI datafile within the GeneticAI folder. It crashes when loading the map each time.
Click to expand...
Sounds like you may have the same issue as Craterus was/is still having.

Some questions:
Have you got the Vanilla Civ 4 1.61 patch?
Do you have Warlords installed? Are you using Vanilla Civ 4 or Warlords?
How do you load the mod?
What map settings are you playing with? (Players, size etc.)
Do you have anything in your custom assets directory/have you modded any of the main game files?

To help solve the problem, it would help if you uploaded some log files for me, so I can try and see where it's crashing. The log files are located in \My Games\Sid Meier's Civilization 4\Logs - there should be about 10-15 of them.

The most useful one for me is GenAIDbg.log, however the others may help. To enable logging change these lines in Civilzation4.ini
Code: 
ShowPythonDebugMsgs = 1
LoggingEnabled = 1
OverwriteLogs = 1
SynchLog = 1
RandLog = 1
MessageLog = 1
 
Hmm, I just came to my compter after having it run all night to find that it was on a game into year 3712. This was one version 0.25. Right when I maximized Civ4 a python error popup came saying something about Domination Victory, but sadly the screenshot I thought I had taken didn't come out. Once I clicked ok to the python then everything looked normal and it was flashing at the bottom of the screen that Washington had just won a space race victory.

I started a new game afterwards and it doesn't appear that the AI values are skewed or anything.

Just thought I'd let you know about that whole thing :)
 
Gunner said:
Hmm, I just came to my compter after having it run all night to find that it was on a game into year 3712. This was one version 0.25. Right when I maximized Civ4 a python error popup came saying something about Domination Victory, but sadly the screenshot I thought I had taken didn't come out. Once I clicked ok to the python then everything looked normal and it was flashing at the bottom of the screen that Washington had just won a space race victory.

I started a new game afterwards and it doesn't appear that the AI values are skewed or anything.

Just thought I'd let you know about that whole thing :)
Click to expand...
You sure it was .25? That bug was supposed to be fixed in .25. I had the same problem with earlier releases but I haven't seen it since. And I certainly haven't seen the game run all the way to 3712 lol.
 
I hope you started the new game with the latest patch. 3712 is quite far into the future - over 1500 turns!

EDIT: Yes, and that would probably be .24 or earlier. Silly bugs.
 
The Great Apple said:
Sounds like you may have the same issue as Craterus was/is still having.

Some questions:
Have you got the Vanilla Civ 4 1.61 patch?
Do you have Warlords installed? Are you using Vanilla Civ 4 or Warlords?
How do you load the mod?
What map settings are you playing with? (Players, size etc.)
Do you have anything in your custom assets directory/have you modded any of the main game files?
Click to expand...

Just to answer your questions whilst I look into log files:

Have you got the Vanilla Civ 4 1.61 patch? Yes
Do you have Warlords installed? Yes Warlords is installed
Are you using Vanilla Civ 4 or Warlords? I'm using Civ 4 to play the game
How do you load the mod? I load the mod through the advanced menu option -> load a mod.
What map settings are you playing with? (Players, size etc.) I have tried a few different numbers and settings
Do you have anything in your custom assets directory/have you modded any of the main game files? I have Sevomod installed but I think that's it. I may just reinstall Civ4, I'm not liking Warlords much anyhow if that is the problem.
 
I really could have sworn that it was .25. It definitely wouldn't be any earlier than .24 since its just playing 8 games per AI. Is there any way to check the version, I can't seem to find one?

I've played most of the games for this set already, I'd say that at least half have 8 games done, but this is the first time that error has come up.
 
Gunner said:
I really could have sworn that it was .25. It definitely wouldn't be any earlier than .24 since its just playing 8 games per AI. Is there any way to check the version, I can't seem to find one?
Click to expand...
Yeah, I really should have versioned them. All future versions will be.

Gunner said:
I've played most of the games for this set already, I'd say that at least half have 8 games done, but this is the first time that error has come up.
Click to expand...
Maybe it was something unrelated. I'll take a look just in case. It would have probably been in your log, but if you started a new game it would have wiped it unfortunetly.

I look forward to seeing those results!
 
Moham said:
Just to answer your questions whilst I look into log files:

Have you got the Vanilla Civ 4 1.61 patch? Yes
Do you have Warlords installed? Yes Warlords is installed
Are you using Vanilla Civ 4 or Warlords? I'm using Civ 4 to play the game
How do you load the mod? I load the mod through the advanced menu option -> load a mod.
What map settings are you playing with? (Players, size etc.) I have tried a few different numbers and settings
Do you have anything in your custom assets directory/have you modded any of the main game files? I have Sevomod installed but I think that's it. I may just reinstall Civ4, I'm not liking Warlords much anyhow if that is the problem.
Click to expand...
All looks good. The Warlords install shouldn't be a problem. Hopefully the logs will help.
 
Here is the GenAIDbg.log from one crash

Code: 
[284135.906] iPlayer: 0, iAI: 0
[284135.906] iPlayer: 0, iAI: 0, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 1
[284135.906] iPlayer: 0, iAI: 1, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 2
[284135.906] iPlayer: 0, iAI: 2, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 3
[284135.906] iPlayer: 0, iAI: 3, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 4
[284135.906] iPlayer: 0, iAI: 4, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 5
[284135.938] iPlayer: 0, iAI: 5, iVarKey: 0, lResult: 0

That was just a standard game set-up.
 
The Great Apple said:
Maybe it was something unrelated. I'll take a look just in case. It would have probably been in your log, but if you started a new game it would have wiped it unfortunetly.
Click to expand...
Yeah, that just occured to me right after I started the game. One of those moronic moments (or in the words of Homer Simpson "doh!" :D) I had just assumed that my screen shot would work, but it didn't.
 
The Great Apple said:
I'd guess the parameters [enabling effective early Axeman rushes] that control how many CITY_ATTACK AIs a player builds in relation to other AIs, how likely a player is to go for bronze working (lots of factors there), and how likely for the AI to declare war early. They'd all have to go up - I don't know in what ratios, but if one of these AIs hits about the right ratio and performs one successfully it should score highly, and new AIs should breed off it to perfect the tactic.

Of course, there would then probably be a strain of AIs that breeds to defend against an early axeman rush, such a rush may become no longer viable, and the axman rush strain may die out.... and then the defensive strain dies and the axeman rush strain comes back up again. This is alot of generations in though.
Click to expand...
There are parameters affecting how likely it is to declare war early?

I think that uniformly aggressive CITY_ATTACK building would create an arms race rather than effective rushes. Is your intention in the end to manually analyze, select, and deploy a set of complementary and balanced parameter sets from among those bred?

But I'd think that effective rushes require more... Assessing your current and near-future military tech advantage. The distance to your neighbor. The chance of your neighbor having or getting effective countering units and city defenses. Your neighbor's likely military production during your buildup. The flexibility to switch to near-maximum unit production, but only in specific circumstances, and without overemphasizing unit production at other times. The willingness to attack your neighbor even if you otherwise would like him.
 
Moham said:
Here is the GenAIDbg.log from one crash

Code: 
[284135.906] iPlayer: 0, iAI: 0
[284135.906] iPlayer: 0, iAI: 0, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 1
[284135.906] iPlayer: 0, iAI: 1, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 2
[284135.906] iPlayer: 0, iAI: 2, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 3
[284135.906] iPlayer: 0, iAI: 3, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 4
[284135.906] iPlayer: 0, iAI: 4, iVarKey: 0, lResult: 0
[284135.906] iPlayer: 0, iAI: 5
[284135.938] iPlayer: 0, iAI: 5, iVarKey: 0, lResult: 0

That was just a standard game set-up.
Click to expand...

Mine looked similar - sent you some dumps TGA! Hope it helps!
 
Jet said:
There are parameters affecting how likely it is to declare war early?
Click to expand...
Probably not directly, but things such as small power advantages and suchlike. It's the sort of thing I'm planning on adding at a later date - at the moment we're altering the default AI, but I'm going to slowly phase in alternate modules to run.
Jet said:
I think that uniformly aggressive CITY_ATTACK building would create an arms race rather than effective rushes. Is your intention in the end to manually analyze, select, and deploy a set of complementary and balanced parameter sets from among those bred?
Click to expand...
The final product may either be one AI, or several which are picked between at the start of the game, probably tailored to certain leader traits as suggested earlier.
Jet said:
But I'd think that effective rushes require more... Assessing your current and near-future military tech advantage. The distance to your neighbor. The chance of your neighbor having or getting effective countering units and city defenses. Your neighbor's likely military production during your buildup. The flexibility to switch to near-maximum unit production, but only in specific circumstances, and without overemphasizing unit production at other times. The willingness to attack your neighbor even if you otherwise would like him.
Click to expand...
Very true. As above, this is the sort of stuff I plan to program in.
 
The Great Apple said:
Ta Moham. I think I might know what is wrong. Attached is a file which I hope will fix the problem.

Unzip in Mods/GeneticAI/Assets/Python/CUSTOM
Click to expand...

I gave this a shot - it appears to be hung on the initialization screen (about an inch of greenbar) after trying to start a standard game - approx 20 minutes hung... amd64, ati graphics card, 2gb of ram

ps - genai file looks similar to above... just sent you two dumps to check out
ps2 - are you recording what settings the ai's are learning on? A small map with 18civs vs a small map with 4 civs will affect results right?
 
Well that's at least an improvement. It shouldn't stay on that screen for much more then a minute. Sounds like it's stuck in a loop.

How do the logs look? Does genAIDbg.log still look the same?

I'll be releasing another fix soon. I've fiddled around with the AI loading script a bit, and it sounds like that's the bit which has gone wrong.
 
You must log in or register to reply here.

Similar threads

T
AI League 2.0
Replies
12
Views
522
Thrasybulos
T
T
The AI Horatius League
2 3
Replies
41
Views
2K
Keler
Keler
M
here's a fair challenge for deity players (difficulty, balance)
Replies
1
Views
869
Morgawr
Morgawr
ATARi144kHz
Highly demanding Civ VII w/o DLC save is needed for Intel/AMD processors reviews
Replies
4
Views
623
Berrern
Berrern
Gen.Washington
Has anyone seen the AI use air units?
Replies
13
Views
671
uppi
U
Back
Top Bottom