This research introduces a novel approach using Large Language Models to automatically generate optimization heuristics that balance multiple objectives simultaneously.

• Creates heuristics that are not only effective but also efficient and scalable
• Applies multi-objective evolutionary algorithms to evolve LLM-generated code
• Demonstrates superior results on classic optimization problems like Bin Packing
• Produces human-interpretable heuristics that can generalize to larger problem instances

For engineering applications, this breakthrough means automated development of optimization algorithms that consider real-world constraints like computation time and code complexity, not just raw performance.

Multi-objective Evolution of Heuristic Using Large Language Model