(Translated by https://www.hiragana.jp/)
GitHub - mascaretti/mcts: Monte Carlo Tree Search - A C++ MPI implementation
Skip to content

Monte Carlo Tree Search - A C++ MPI implementation

License

Notifications You must be signed in to change notification settings

mascaretti/mcts

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Monte Carlo Tree Search

Implemented by

(Source code available on GitHub)

Requirements

  • C++11 compiler (tested with g++ and clang)
  • MPI (for parallelization)

Optional:

Compiling

  • Run $ make [options] in the main directory.
    • Selecting DEBUG=yes (DEBUG=no) enables (disable) debugging (with GDB, for instance)

Running

  • Parallel:

    • run $ mpiexec -np [number of cores] [executable] to run the parallel scripts.
  • Speedup test:

    • run test_gen.sh to run all the tests in the folder speedup_test.
  • Serial:

    • run $ ./[executable] to execute the serial scripts.

Description

  • src: contains the source code and the template classes that implement both the games and the search tree
  • examples: contains examples of matches, namely ai vs. random player, ai vs. ai and ai vs. human player
  • test: contains various tests concerning data structures (struct_test), games (nim_test and oxo_test), parallel speedup (speedup_test) and data analysis (analysis)

Developed and tested on Debian, CentOs and MacOS X