Jump to content

  • Log In with Google      Sign In   
  • Create Account

How to set up a audimentary AI?


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
5 replies to this topic

#1 Ironmaggot   Members   -  Reputation: 221

Like
0Likes
Like

Posted 19 December 2013 - 09:03 AM

How to setup a rudimentary AI for my game?

 

I want to set up a starting block for my AI, so that I get some basic functionality. I just got to the stage, where implementing a rudimentary AI is now one of my next steps. I got basic framework built in my game and I need something to "populate" the game.

 

My game is a strategy game. It has got stars that each have their own coordinates, fleets that orbit around the stars and fleets can travel between stars. But fleets can only travel from star to star. Each star has a maximum travelling range. Fleet on the star can only travel to stars that are within the max travelling range.

 

I want to create an AI for non-player empire. That AI would have to at least move around its fleets randomly (move them around in its own territory and/or move them onto mine). I plan on making the AI variable based eventually.

 

Which means that each empire has personality and stances towards other empires. Personalities include agressiveness-passiveness, honor-deception and etc. Stances towards other empires include disposision, balance-of-power, etc. So for example agressive empire attacks those whose military power is smaller than theirs and also attacks those whose power rating is rapidly rising, to eliminate possible new threats. Honorful AI would require justification to launch campaigns, meaning that it will not attack anyone without provocation. And honorful empire almost never will end any pact's it has made and will never form pacts with empires that they have low disposision twards to. But decietful empire will form any pack that is of use to them. So that decietful empire might form a military-acess pact with the enemy of his ally should that benefit it more.

 

But the above won't be added in any time soon. So, give me pointers to make a rudimentary AI and if possible, in a way that greates a small stepping stone towards my goal.

 

And any links or reccomendations about further learning about AI are welcomed. Or if you think that I could pull off similar result eventually in a more effective way, you are welcome to share.



Sponsor:

#2 Álvaro   Crossbones+   -  Reputation: 13912

Like
1Likes
Like

Posted 19 December 2013 - 09:12 AM

Random choices seems like a good place to start.

struct Agent {
  Action pick_action(std::vector<Action> const &actions) {
    return actions[std::rand() % actions.size()];
  }
};

Now write whatever code you need to in order to make this AI work. You need to be able to determine what the AI's available actions are and how to execute them.

 

Whenever you want to replace this dummy AI with something more reasonable, you may want to pass a view of the game state as well, or you may want to beef up the `Action' class so it contains enough information for the agent to decide which action is his favorite. This will naturally lead you to a utility-based AI architecture, which is my favorite paradigm for AI in general, and which lends itself very well to implementing personalities.


Edited by Álvaro, 19 December 2013 - 09:13 AM.


#3 Ironmaggot   Members   -  Reputation: 221

Like
0Likes
Like

Posted 19 December 2013 - 09:52 AM

Random choices seems like a good place to start.

struct Agent {
  Action pick_action(std::vector<Action> const &actions) {
    return actions[std::rand() % actions.size()];
  }
};

Now write whatever code you need to in order to make this AI work. You need to be able to determine what the AI's available actions are and how to execute them.

 

Whenever you want to replace this dummy AI with something more reasonable, you may want to pass a view of the game state as well, or you may want to beef up the `Action' class so it contains enough information for the agent to decide which action is his favorite. This will naturally lead you to a utility-based AI architecture, which is my favorite paradigm for AI in general, and which lends itself very well to implementing personalities.

 

I can understand your point, but I can't understand your code. I only know Python, so can you explain that code a bit?



#4 Waterlimon   Crossbones+   -  Reputation: 2638

Like
1Likes
Like

Posted 19 December 2013 - 10:23 AM

 

Random choices seems like a good place to start.

struct Agent {
  Action pick_action(std::vector<Action> const &actions) {
    return actions[std::rand() % actions.size()];
  }
};

Now write whatever code you need to in order to make this AI work. You need to be able to determine what the AI's available actions are and how to execute them.

 

Whenever you want to replace this dummy AI with something more reasonable, you may want to pass a view of the game state as well, or you may want to beef up the `Action' class so it contains enough information for the agent to decide which action is his favorite. This will naturally lead you to a utility-based AI architecture, which is my favorite paradigm for AI in general, and which lends itself very well to implementing personalities.

 

I can understand your point, but I can't understand your code. I only know Python, so can you explain that code a bit?

 

 

The object Agent (the AI) is given a pile of possible actions it can take, out of which a random one is chosen.


o3o


#5 wodinoneeye   Members   -  Reputation: 877

Like
0Likes
Like

Posted 14 January 2014 - 11:48 PM

Do a design on paper of basic strategies for different  NPC sides to make use of

 

and then from those playing strategies come up with basic tactic matching that way of playing.

 

 

High risk vs low risk   (overwhelming advantage vs take a risk minimum resource for the job)

all offensive vs mostly defensive (grab to grow vs defend strongly what you take

Go for the gold or grab everything  (solid front vs narrow targeting of best resources)

depending on game mechanics -  low tech vs high tech ,    production development/improvement vs cheapest production

short term planning (reevaluate and change) vs long term (following thru)

Localized (flexibility) vs everything following same strategy (cohesiveness)

 

(note all of the above have degrees of how far they lean into the startegy)

 

many of the above are interrelated and others independant  (allowing mix and match)

 

 

the tactics to implement the above generalized strategies are much more game mechanics dependant so I cant really list any


--------------------------------------------Ratings are Opinion, not Fact

#6 IADaveMark   Moderators   -  Reputation: 2531

Like
0Likes
Like

Posted 15 January 2014 - 11:56 AM

Ask yourself how/why/when YOU would make a decision. Figure out which pieces of information are important to that decision. Get that information and process so that you can make the decision.

 

This is how all AI is done regardless of language or structure.


Dave Mark - President and Lead Designer of Intrinsic Algorithm LLC

Professional consultant on game AI, mathematical modeling, simulation modeling
Co-advisor of the GDC AI Summit
Co-founder of the AI Game Programmers Guild
Author of the book, Behavioral Mathematics for Game AI

Blogs I write:
IA News - What's happening at IA | IA on AI - AI news and notes | Post-Play'em - Observations on AI of games I play

"Reducing the world to mathematical equations!"




Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS