Artificial Intelligence for MOBA games

I am just getting started in learning AI for game development.

I was wondering what kind of algorithms are implemented in developing the AI in MOBA games (such as DotA, LoL, Infinite Crisis).

Also, does anyone have any idea about the number of developers in a team required for the development of such an AI?

 

If you are just starting out, forget about the big guys and focus on the basics. Riot probably has a full team working on the A.I as they have ton of champions which all behave very differently which requires fair bit of effort in their part.

I recommend you to check out "programming Game AI by example" by Mat Buckland which is probably one of the best books out there that goes trough many core principles with great depth. I loaned mine from a colleague but you can probably order it, or get a pdf version trough official (or unofficial) channels. Few pages seem to be available on google books as well.

If you just want pointers, then here's a few key words for Google:
- Fixed state machines
- steering behaviors
- Path finding (a*star)
- Fuzzy logic
- Goal-Driven Agent Behavior

Most if not all these are used in pretty much every game that requires more from a.i than simple pursue behavior.

 

Thanks for your reply Vipsu. I think that book will be a good starting point for me.

Btw, I am not looking to create a MOBA right away, of course. I was just curious as to how many people work on such a project (AI part).

 

Hey, just wanted to note that most MOBA games use behaviour trees. You can find some articles of Riot Games talking about it in aigamedev.com

I myself developed an AI for MOBA, you can see both articles here:
On the Development of Intelligent Agents for MOBA Games: http://homepages.dcc.ufmg.br/~vnsilva/papers/147511_2.pdf
A tutor agent for MOBA games: http://homepages.dcc.ufmg.br/~vnsilva/papers/147512_1.pdf

 

6 years later but I have to try... The links are not working, do you have those papers so far? If yes, can you share them?

Thanks