Sim2Schedule: A Simulator-Guided LLM Framework for Autonomous Open-Pit Mine Scheduling

The paper presents Sim2Schedule, a simulator-guided LLM framework for autonomous open-pit mine scheduling that integrates a custom simulator to guide decision-making. This approach allows the LLM to generate interpretable extraction and processing schedules in a zero-shot, data-secure environment, achieving 94% to 99% recovery of optimal Net Present Value (NPV) from a newly developed Mixed-Integer Linear Programming (MILP) formulation while scaling linearly in computation time. This framework offers a practical alternative to classical optimization methods, addressing the challenges of dynamic industrial environments with complex constraints.