Jump to content

  • Log In with Google      Sign In   
  • Create Account

Banner advertising on our site currently available from just $5!

1. Learn about the promo. 2. Sign up for GDNet+. 3. Set up your advert!


Member Since 02 Jan 2002
Offline Last Active Yesterday, 01:10 PM

#5221331 AI Simulation

Posted by IADaveMark on 04 April 2015 - 09:02 AM

How about doing one of those things at a time instead of trying to combine them all into one master project?

#5221329 Design a game that is good for a neural network

Posted by IADaveMark on 04 April 2015 - 09:00 AM

I agree with everyone else that you are:


  1. Doing it backwards by choosing a game to match the algo
  2. Putting the word "good" in the wrong place... i.e. "a game that is good for a neural network". In theory, you could have a very, very simple game that NN could solve. However, it would likely be a pathetic game. It is very likely you will achieve the different goal of designing "a good game that a neural network can solve."

If you are just trying to learn how to do a NN, don't design a game. Use something standard.


When you are done, put NNs behind you and concentrate on doing AI with other algos that have far more application in game AI.

#5216393 How to make bot playing game with neural net

Posted by IADaveMark on 13 March 2015 - 11:44 PM

Or just use a freaking influence map and be done with it.

#5212380 "Frontline"

Posted by IADaveMark on 22 February 2015 - 09:44 PM

It's called "influence mapping". I believe the article was by Steve Woodcock in AIGPG1. There have been others since then in a variety of places (including one in the upcoming AI Pro 2 written by myself).

#5192645 Algorithm Suggestion to implement AI in a 2D soccer/football game

Posted by IADaveMark on 13 November 2014 - 07:50 AM

Mat Buckland has a pretty nifty soccer game implementation as part of the examples in his book, "Game AI by Example".

#5188538 Neural net for game state prediction

Posted by IADaveMark on 22 October 2014 - 10:12 AM

  1. This sounds a lot like a homework problem (largely because it has absolutely no practical use). We don't help people with homework here. Against the rules.
  2. Use math instead (largely because using a NN here has no practical use). If your teacher gets snotty saying that wasn't the assignment, tell said teacher, "I decided that it would be preferable to use the right tool for the job." See if you get extra credit.

#5187573 Choosing a Machine learning method for a game.

Posted by IADaveMark on 16 October 2014 - 09:30 PM

Design an AI that plays the game first. Until you do that, there is no "learning" that can happen.

#5184865 Using machine learning to predict the collapse & stabilization of complex...

Posted by IADaveMark on 03 October 2014 - 06:27 PM

Your best bet is to design the simulation with rubber-banding constraints so that the farther certain things get out of whack the less likely they are to proceed further in that direction. That's how many large scale decision systems are anyway.

#5184647 Using machine learning to predict the collapse & stabilization of complex...

Posted by IADaveMark on 02 October 2014 - 03:19 PM

If you have access to professors and math grad students, find one with a decent grasp on chaos mechanics and have them explain to you why you can't do what you're attempting to do.


There's probably an "n" in their answer.


For a shorter solution, look up Poincare's "3 body problem".

#5184554 Using machine learning to predict the collapse & stabilization of complex...

Posted by IADaveMark on 02 October 2014 - 06:50 AM

OK... so it sounds like you are trying to figure out which butterfly is going to cause the hurricane -- i.e. "solve chaos theory". Yeah... that's gonna be a bitch. Even if you just run the simulation looking for values that start to arc in one direction, you don't truly know how far it will get until it starts to correct.

#5183308 A question about behavior trees

Posted by IADaveMark on 27 September 2014 - 10:21 AM

You did hit on the correct solution in that the behavior is not "move to A" or "move to B" -- but rather "move in the direction of A/B". In fact, it is likely better to have a single node that is "move toward current long term destination" and have that destination be set elsewhere. Regardless, when the bandits attack, they trigger all the higher priority defense behaviors. When those are no longer valid (e.g. no more bandits), the highest remaining behavior is "move in the direction..." and the caravan resumes.


The actual switching of destinations would be a higher priority behavior than the "move in the direction of" that has as its condition "have I arrived?". If you have arrived at the specified destination, you would then run whatever logic it would be for selecting a new one (even if it is just flipping A/B). Now, since you are no longer at your specified destination, "move towards" is the highest ranking behavior and off you go.

#5180084 The difference between Logic, Reasoning, and Thinking; Data, Information, and...

Posted by IADaveMark on 13 September 2014 - 10:02 AM

Oh brother...

#5177320 Algorithmic Ecology: Machine Learning AI Engine

Posted by IADaveMark on 31 August 2014 - 07:10 PM

At least you admit that it is for personal reasons... because it is the biggest pariah in game development.


I haven't looked in much detail but it seems to me that in order to get something that would be specific the game design you wanted it to be for, you would have to put in a lot of time just to put together the training set. And even then, you have no idea what you are going to get out of it. Seems like a real blindfolded dart-throwing approach.

#5177289 Algorithmic Ecology: Machine Learning AI Engine

Posted by IADaveMark on 31 August 2014 - 04:42 PM

Why machine learning?

#5175302 How to make a AI choose?

Posted by IADaveMark on 21 August 2014 - 10:32 AM

"Utility" > "fuzzy"



I approve of this.