BPP-Search: Enhancing Tree of Thought Reasoning for Mathematical Modeling Problem Solving
This work addresses a bottleneck in reinforcement learning for mathematical modeling in operations research by providing better datasets and methods, though it is incremental as it builds on existing tree-of-thought and reinforcement learning techniques.
The authors tackled the problem of limited annotations in open-source operations research datasets by releasing the StructuredOR dataset with comprehensive labels for mathematical modeling processes, and proposed BPP-Search, an algorithm that integrates reinforcement learning into tree-of-thought reasoning, which significantly outperformed state-of-the-art methods in accuracy and efficiency on multiple datasets.
LLMs exhibit advanced reasoning capabilities, offering the potential to transform natural language questions into mathematical models. However, existing open-source datasets in operations research domain lack detailed annotations of the modeling process, such as variable definitions, focusing solely on objective values, which hinders reinforcement learning applications. To address this, we release the StructuredOR dataset, annotated with comprehensive labels that capture the complete mathematical modeling process. We further propose BPP-Search, an algorithm that integrates reinforcement learning into a tree-of-thought structure using Beam search, a Process reward model, and a pairwise Preference algorithm. This approach enables efficient exploration of tree structures, avoiding exhaustive search while improving accuracy. Extensive experiments on StructuredOR, NL4OPT, and MAMO-ComplexLP datasets show that BPP-Search significantly outperforms state-of-the-art methods. In tree-based reasoning, BPP-Search excels in accuracy and efficiency, enabling faster retrieval of correct solutions. The StructuredOR dataset is available on Huggingface https://huggingface.co/datasets/LLM4OR/StructuredOR and GitHub https://github.com/LLM4OR/StructuredOR.