Edited for brevity.
I made a post 8 years ago about this Stratego-like game: https://www.gamedev.net/forums/topic/585572-game-of-the-generals-ai/. Are there advances in AI in the last eight years you recommend?
As I understand, MCTS/UCT/AlphaZero is not applicable because of hidden information.
My AI implementation uses MC playouts + UCB1 + inference of opponent's pieces. It's not effective, though it makes smart moves on the end game, like the flag piece running away.
Bayesian inference. Not new, but the go-to tech for hidden information.
Dave Mark - President and Lead Designer of Intrinsic Algorithm LLC
Professional consultant on game AI, mathematical modeling, simulation modeling
Co-founder and 10 year advisor of the GDC AI Summit
Author of the book, Behavioral Mathematics for Game AI
Blogs I write:
IA News - What's happening at IA | IA on AI - AI news and notes | Post-Play'em - Observations on AI of games I play
"Reducing the world to mathematical equations!"
On 9/4/2018 at 10:42 PM, IADaveMark said:Bayesian inference. Not new, but the go-to tech for hidden information.
Thank you for your reply. I think this is similar to what I've done for piece prediction: I keep track of piece's possible highest and lowest ranks after challenges. I then use this information to generate plausible board configuraions.
Currently reading this paper on Information Set Monte Carlo Tree Search: https://pure.york.ac.uk/portal/files/13014166/CowlingPowleyWhitehouse2012.pdf
Longer version: https://pdfs.semanticscholar.org/d10e/31ed85cc6ea79d3d961730da2b07c32aa984.pdf
Hello, I implemented a cheating MCTS. Definitely better, but still beatable. Will improve on that before doing anything fancy.
