# InstructMPC: A Human-LLM-in-the-Loop Framework for Context-Aware Control

arXiv: https://arxiv.org/abs/2504.05946 · HTML: https://mlsanity.com/paper/2504.05946/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Apr 2025.

## Built on by (most cited first, 3 of 3)

- [Bridging Natural Language and Microgrid Dynamics: A Context-Aware Simulator and Dataset](https://mlsanity.com/paper/2604.05429/) (2026)
- [Agentic MPC for Semantic Control System Resynthesis](https://mlsanity.com/paper/2606.12774/) (2026)
- [LLM-Based Hierarchical Coordinated Control with Continuation-Aware Policy Learning](https://mlsanity.com/paper/2608.15041/) (2026)

## Abstract (arXiv)

Model Predictive Control (MPC) is a powerful control strategy widely utilized in domains like energy management, building control, and autonomous systems. However, its effectiveness in real-world settings is challenged by the need to incorporate context-specific predictions and expert instructions, which traditional MPC often neglects. We propose InstructMPC, a novel framework that addresses this gap by integrating real-time human instructions through a Large Language Model (LLM) to produce context-aware predictions for MPC. Our method employs a Language-to-Distribution (L2D) module to translate contextual information into predictive disturbance trajectories, which are then incorporated into the MPC optimization. Unlike existing context-aware and language-based MPC models, InstructMPC enables dynamic human-LLM interaction and fine-tunes the L2D module in a closed loop with theoretical performance guarantees, achieving a regret bound of O(√Tlog T) for linear dynamics when optimized via advanced fine-tuning methods such as Direct Preference Optimization (DPO) using a tailored loss function.