Researchers have developed a method called AutoOR to help large language models (LLMs) automatically translate complex optimization problems into a format that can be solved by computers. This is achieved through a combination of synthetic data generation and reinforcement learning. The approach has been tested on several established benchmarks in the field of operations research, with promising results. In particular, it was able to match or outperform larger models on some tasks, even when trained on limited initial data. --- Why it matters: This work is significant for engineers and researchers working on AI applications in industrial settings, as it provides a potential solution to the challenge of scaling up the use of LLMs in optimization problems. By enabling LLMs to automatically translate complex descriptions into solver-ready formulations, AutoOR can accelerate decision-making in industries such as manufacturing and logistics. Source: https://arxiv.org/abs/2604.16804