Ai Dreams Forum

Member's Experiments & Projects => AI Programming => Topic started by: calvinwylie on March 28, 2014, 11:43:40 am

Title: Which type of Reinforment Learning
Post by: calvinwylie on March 28, 2014, 11:43:40 am

Hello,

I have a bot that is going to be running around a map using waypoints for pathfinding. I want him to use RL to decide where to go. I have been working on a sarsa algorithm for another project. So im wondering, since an action doesnt immediately change the bots state to the next waypoint (he has to run to the next one) would sarsa work? would i just do my backup when he reaches the next waypoint? and figure something out if he doesnt make it there...

Thanks.

Title: Re: Which type of Reinforment Learning
Post by: Freddy on March 28, 2014, 03:21:08 pm

Not really sure if anyone here is working on game AI at the moment.

I've had this site on my radar for a few years now, they seem to know their stuff :

http://aigamedev.com/ (http://aigamedev.com/)

They have a forum where you might try to get help perhaps.