MCTS Simulation Phase

You are executing the SIMULATION (rollout) phase of Monte Carlo Tree Search.

LLM as Heuristic Policy

Use your knowledge to:

1. •Guide the rollout toward realistic outcomes
2. •Evaluate terminal states with meaningful scores
3. •Detect dead ends early to save computation

Simulation Algorithm

1. •Start from the expanded node
2. •Rollout to terminal state:
   • •Select actions using LLM policy (not random!)
   • •Simulate state transitions
   • •Continue until terminal or max depth
3. •Evaluate the outcome:
   • •Success: positive reward (e.g., 1.0)
   • •Partial success: proportional reward (e.g., 0.5)
   • •Failure: zero or negative reward

Using MCP Tools

Call mcts_simulate with:

• •node_id: The node to simulate from
• •max_depth: Maximum rollout depth (default: 10)
• •evaluation_criteria: What constitutes success

The tool returns:

• •terminal_state: The final state reached
• •reward: Numerical evaluation [0, 1]
• •rollout_path: Sequence of actions taken
• •reasoning: Explanation of the evaluation

Simulation Strategy

For the current context: $ARGUMENTS

Rollout Policy

Instead of random rollout, use informed policy:

1. •At each step, consider 2-3 likely actions
2. •Choose based on domain knowledge
3. •Prefer actions that lead to decisive outcomes

Evaluation Criteria

For Research:

• •Does the path lead to valid conclusions?
• •Is evidence sufficient and reliable?
• •Are there logical gaps?

For Planning:

• •Does the plan achieve the goal?
• •Are resources within budget?
• •Are there critical risks?

For Coding:

• •Does the solution work correctly?
• •Is the code clean and maintainable?
• •Are edge cases handled?

Reward Assignment

reward = completeness * correctness * efficiency

Where each factor is in [0, 1]:

• •completeness: How much of the goal is achieved
• •correctness: How valid is the solution
• •efficiency: How elegant/optimal is it

Output

After simulation, report:

1. •Terminal state reached
2. •Reward value with breakdown
3. •Key insights from the rollout
4. •Any observations to record

Proceed to BACKPROPAGATION with the reward.