Experimental Ai Algorithms made here?

[MORON]

Well, one of the biggest complaint about Civ is about its less than human level Ai. Since this forum has collected some of the best Civ players around, I guess this would be an good place to come up with some algorithms for the Ai.

So any clever algorithms that can fix obvious problems with the ai?

It is an long shot, but it is still an better shot at getting an better game than repeating stuff from the Civ3 list IMO.

[Max Sinister]

If you or someone else were going to program a better AI algorithm for Civ, there are many questions to answer. Questions like these:

- Where do you build a city, if you've got a settler?
- How many of your resources do you use for building settlers / units / buildings / small and great wonders? Which ones? And if you know: Where do you build what? (A harbor at the sea, a great wonder in the biggest city, a factory in the mountains with mines, units in cities near the enemy, settlers in the cities near the unsettled wilderness...)
- Which other civ do you make peace with, and which one do you attack? Where do you attack them - units or cities, and which ones?

[Urban Ranger]

Originally posted by Max Sinister
- Where do you build a city, if you've got a settler?

I think it will be better if the ability to build city at will is removed. Instead, a city will appear when there's sufficient population on a tile, with population growth controlled by a new model.

[Adagio]

Originally posted by Urban Ranger
I think it will be better if the ability to build city at will is removed. Instead, a city will appear when there's sufficient population on a tile, with population growth controlled by a new model.

I have thought about something like that for some time, but I believe hell freezes over before Firaxis implements that idea, since it's not Civ'ish

[Jon Shafer]

Something like this should be possible using scripting, if Firaxis allows as much moddable as they're hinting at.

[Adagio]

For the time being I don't believe they'll make it that moddable... I believe they'll just make most of the variables used in the game moddable (like RoN), not as moddable as ctp where you could actually do some scripting yourself...

But I wouldn't mind getting suprised, I just don't want to get my hopes up...

[Jon Shafer]

XML is used for storing data (i.e. your variable modding).

What do you think they're using Python for?

The announcement a while ago mentioned that the AI will be moddable with Python. If things so far-reaching as the AI's behavioral code are moddable then this should not be an issue.

[Adagio]

Ok, I haven't heard anything about how modable it will be or what they will use for it... haven't kept up with the news about Civ4... Maybe I should start reading about it before I comment on it

[Urban Ranger]

XML? Isn't that a bit overkill?

[Dale]

AI modding and Model modding are two completely different things. You must have both for it to work. A modded Model will not work if the AI is not modded to handle it, and vice-versa.

It's one thing to say "a city will appear when there's sufficient population on a tile" as a scripted MODEL but then it's a completely different things to have the scripted AI to handle it. (See my CTP2 script Colonisation which actually did this!)

So with that in mind, Firaxis saying that the AI is moddable by Python does not mean you can do Model modding. As much as I wish it will be true I can't see them doing it. To me it sounds more like Personality modding rather than actual AI modding.

Dale

[Max Sinister]

Originally posted by Urban Ranger
I think it will be better if the ability to build city at will is removed. Instead, a city will appear when there's sufficient population on a tile, with population growth controlled by a new model.

Sorry Ranger, I wouldn't like it. In our world kings have founded many cities, villages and colonies, and *they* decided where to put them. It's ok if the pop. model decides which city will grow faster, but I'd like to decide where to put them myself. If I should choose the wrong places, well, I might lose then. So be it.

[MORON]

If only we have an measure of "power" in calculating the effects of each decision the ai makes, and maximize the power of Ai races.

*uses some economic theories to model civ decisions*

I'm thinking
War potential = production x war tech
Using this, we can balance production with tech tree. It shouldn't be to hard to find the maximum delta war potential using this equation.

War capacity = units x war tech
War capacity should be maintained to reach parity with humans if possible, to prevent invasions.

The ai should first try to maintain parity militarily, than any surplus in production after the two would be invested in production and tech to maximize future war making ability. There should also be an "risk" function on how far ahead the ai looks at, and it should change depending on how endangered the Ai is. If an Ai is on a island away from all other players, it can take high risk, long term maximum as the goal.

[Urban Ranger]

Originally posted by Max Sinister
Sorry Ranger, I wouldn't like it. In our world kings have founded many cities, villages and colonies, and *they* decided where to put them. It's ok if the pop. model decides which city will grow faster, but I'd like to decide where to put them myself. If I should choose the wrong places, well, I might lose then. So be it.

Well, while I am not sure on the mechanics of how these kings founded cities (some references will be nice), I don't think they created Settler units and put them in the middle of nowhere.

[alms66]

Originally posted by Max Sinister Sorry Ranger, I wouldn't like it. In our world kings have founded many cities, villages and colonies, and *they* decided where to put them. It's ok if the pop. model decides which city will grow faster, but I'd like to decide where to put them myself. If I should choose the wrong places, well, I might lose then. So be it.

This is true, however, as UR points out, they didn't move settlers into place to found the city, they merely said, "Hey you, build a city here." Actually, he sent workers to build the infrastructure. And since the workers and their families were already there, you had a small amount of people to live there already, while the rest came from the surrounding countryside.

I would say that the population founds their own cities, but the ruler still has the ability to found some cities as well, though it should be expensive and/or time-consuming. The method for founding could be as simple as paying a large cash amount for it and having workers build it (along with the cash deposit). The survival of a city the ruler founds is dependent upon its placing, just as every other city. Though, from what I know, most cities attract people, unless they run out of water.

[Urban Ranger]

Originally posted by alms66
I would say that the population founds their own cities, but the ruler still has the ability to found some cities as well, though it should be expensive and/or time-consuming. The method for founding could be as simple as paying a large cash amount for it and having workers build it (along with the cash deposit).

That's a good idea. Let the player to designate a square with certain amount of people already in it as a City Site. There will be a short period of population boom (say, cash rewards is given to people moving to the city), and the among of population increase is dependent on the surrounding areas and how high the original population level is.