mcts

Monte Carlo Tree Search for Causal Intervention Planning.

This module implements a canonical MCTS algorithm (Select, Expand, Rollout, Backpropagate) for finding minimal causal interventions that transform a current symbolic state into a goal state. The algorithm operates purely on symbolic representations without requiring simulation.

Author: Yazz Warsame

Functions

load_scenario(scenario_name)	Load scenario configuration and initial state from JSON files.
main()	Main entry point for running MCTS intervention planning.

Classes

CausalMCTS(initial_state, goal_state, ...)	MCTS-based planner for causal intervention planning.
MCTSNode([interventions, parent])	Represents a node in the MCTS tree.