Advertisement

AI for strategy game

Started by October 17, 2010 01:59 AM
6 comments, last by Azul 14 years ago
Hello.
I need help in creating AI for strategic games.
The main difficulty lies in the fact that it is not just a strategy, a global strategy, ie the player controls the whole country ...
Game I try to do so micromanagement as little as possible ...

Please play my game and tell me what you saw AI. Or even be your own strategy ...
What are you doing?
Why are you doing?
Etc.

My game: http://isgameplanet.at.ua/load/3-1-0-4
My game: isgameplanet.at.ua
You might want to translate that to english... the majority of people that will read this post are probably not going to understand a single thing on that site. My self included.
[ dev journal ]
[ current projects' videos ]
[ Zolo Project ]
I'm not mean, I just like to get to the point.
Advertisement
Sorry, but I'm not very fluent in English ... :) So do not expect high-quality translation ...
Now I can offer you translations of the text, but to embed them in the game can only a week or more. I just had almost finished the next assembly, and the spread is not one assembly does not want.

Here kratenky translate all texts from the game:
Ход игрока 1, ход 1 - Player's 1, turn 1
Население - Population
Рекруты - Recruits
Рост населения - Population growth
Рекрутский набор - Recruitment
Доход - Income
Армия - Army
Время волнений - The riots
Шанс исследовать - Chance to explore
Налог - Tax
Рекруты - Recruits
Повысить - Increase
Понизить - Lower

Thank you for your attention;)

PS an interesting case ... Is it hard to learn Russian? We then speaks English as though it ... Are you in Rooskey a word you do not know ... One wise man said, "The more languages you know, the smarter you can consider himself" ... I can speak 3 languages is free and on 3 with some difficulties ... And you? ;)
My game: isgameplanet.at.ua
I don't speak russian yet (I am learning Chinese) but I can recomend general AI methods.

Bruteforce:
http://en.wikipedia.org/wiki/Monte_Carlo_method
There is a russian version but the english one seems to be more relevant to games.

Evolution:
Make a neural network that will play against a mutated copy of itself. It's input senses will be simplified statistics about who has more power in land/sea/air/anti-air on different regions. When it gives a signal, make a high level action coded by hand.
http://en.wikipedia.org/wiki/Neural_network

The most vital part of being good at strategy games is balance between growth, expansion, defence, offence and technology because that is about knowing your enemy. Everything else can be calculated with prooven correctness.
Quote: Original post by Ivin
PS an interesting case ... Is it hard to learn Russian? We then speaks English as though it ... Are you in Rooskey a word you do not know ... One wise man said, "The more languages you know, the smarter you can consider himself" ... I can speak 3 languages is free and on 3 with some difficulties ... And you? ;)


It doesn't matter how many languages other people in this forum speak. You are asking for help in an English-language forum, and you can only expect people here to know English.

Monte Carlo methods have my vote as the first thing to try.

Hi, im researching putting some AI into a orge-battle style game. From reading up on the monte-carlo method, I see how it applies to a game like Battleship, but how would it apply to a moving target. Imagine battle ship but you could move your ship one square each turn.

Sorry if this is a dumb question, AI is new territory to me.
Advertisement
Books:

Intro to Game AI, by Neil Kirby
AI By Example, by Mat Buckland
AI Game Programming Wisdom 1-4, Steve Rabin editor

Web Site:

www.aigamedev.com

I have an article in AI Game Programming Wisdom 3 on the AI we built for Kohan 2, which was a strategy game. That might not be a bad starting point.
Thanks! I'll check those out, this is probably one of the toughest parts of programming a game so far!

This topic is closed to new replies.

Advertisement