Sign in to follow this  
  • entries
    42
  • comments
    61
  • views
    32836

Untitled

Sign in to follow this  

290 views

OK, so I have a simple little environment to play in, now it's time to do something with it.

This first project wasn't meant to be in space but what the heck!

I was flicking through Rufus Isaac's book "Differential games" and that's given me a few ideas about how to proceed. The plan for this project was to provide an environment to experiment with some AI pursuit problems. So what I'm going to begin with it the 'guarding a target' problem. The basic idea is that the player must enter the target region to win and the computer player must intercept the player to win. There is an optimal strategy for the most basic formulation of this problem and I can use that as a benchmark to gauge other results by. Then I want to train a neural network to learn how to intercept, using the benchmark as a measure of the error.

That's all fairly simple stuff but the problems begin to get harder when the player can move faster than the computer player, or is more maneuverible, or there are obstacles that interfere with the optimal strategy, or when there are multiple computer players. There are many ways to extend or complicate the problem. First I need to get the simple case working.

Wish me luck! [smile]




Sign in to follow this  


2 Comments


Recommended Comments

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now