Skip to content

Typescript implementation of simultaneous move Monte Carlo Tree Search with chance events.

Notifications You must be signed in to change notification settings

DiogoRNeves/sm-mcts

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

sm-mcts

Typescript implementation of simultaneous move Monte Carlo Tree Search with chance events using a forward only simulator.

This is a work in progress still in very very very very early stages.

Motivation for this work is to use MCTS to choose the best action in a Pokémon Battle using showdown simulator library without saving/loading states.

The approach for this library is to force an implementation of a simulation (which will be a simulator wrapper for showdown) to be passed into the MTCS. This will allow for a multitude of possibilities on the motivation case, such as:

  • simulation for multiple generations
  • battle state manipulation before start (such as setting weather)
  • include battle team selection (i.e. for hardcore nuzlocke use)
  • include gaunlets with or without healing between battles (i.e. for hardcore nuzlocke use)
  • filtering opponent's moves (i.e. mimic the game's AI)

This will initially be done without paralelization, but if I manage to make it work properly I may try using multiple concurrent simulators.

About

Typescript implementation of simultaneous move Monte Carlo Tree Search with chance events.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published