TASEmulators / TASEmulators/BizHawk

Feature Request: MCTS Bot

Open
#1,764 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Ext. tool: Future tool Request: Feature/Enhancement
Dominant language
C#
Stars
2.8k
Forks
468
PR merge metrics
No merged PRs in 30d

Description

Proposal: Similar to BasicBot, but use a smarter algorithm.

Main bullet points for implementation:

  • Duplicate the BasicBot menus (rename: "Advanced Bot [ unstable ]" )
  • Generate list of possible moves to select from based on selected menu parameters
  • Implement scoring function
  • Implement MCTS with UCT
  • Optional: Disable audio/video, and un-throttle core while bot is running.

Note, e.g. that the new scoring function is required. If you wanted to search for a maximize a RAM value, you'd likely want to score a 0xFF as a 1, a 0x80 as a 0.5, and a 0x00 as a 0. This is roughly equivalent to a max() operation, but importantly keeps scoring independent of the status of the other branches.

Reason for proposal:
Basic bot is useless for all but the most narrow applications. While MCTS is not appropriate for entire TAS, the hope is that it can search small sections of say 30 frames, while discovering novel tricks.

Weaknesses:
MCTS is likely not better at searching for true random events. It is most efficient at searching functions that have a smooth gradient, such as optimizing movement.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by locating BasicBot and its menus, then trace how menu parameters produce possible moves. The proposal calls for an Advanced Bot using a scoring function and MCTS with UCT; done means those listed capabilities work for short searches such as roughly 30 frames, with the optional runtime changes considered separately.

Written by the indexing model from the issue text.

Assessment

Tech stack
csharp
Domain
tooling
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.