Archived

This topic is now archived and is closed to further replies.

Sagar_Indurkhya

Pathfinding and Reasoning

Recommended Posts

Hello. I have been doing some research into the A* Algorithm, and I have been wondering how I could implement a reasoning thing into it. Such as whether approaching from behind is better than approaching from the front? I want to maybe let the computer find out for itself? I want one ai model to be able to adapt itself in a different environment. I have thought that maybe letting the computer do thousands of simulations against itself, recording the events of what happens, and letting it weight the different decisions it can make based on how successful they are. I have a haunch that doing 1000''s of simulations isn''t going to work right. Any ideas on this would be appreciated. Sagar.Indurkhya

Share this post


Link to post
Share on other sites
You mean like giving the nodes behind the enemy a lower cost than in front? I thought about it too. But I talked to a psychology guy, and he said there are two kinds of AI. One is like Deep Blue, simply referencing a database, and one uses reason based on experiance. And the experiance only comes from trying things for yourself. So would I let the different bots try to calculate their own heuristic scales? Sounds good to me!!

Sagar Indurkhya

Share this post


Link to post
Share on other sites