Search papers, labs, and topics across Lattice.
This paper explores the use of Large Language Models (LLMs) to automate the reformulation of constraint models in combinatorial problem-solving, aiming to enhance solver efficiency. By employing an evolutionary search framework inspired by Automatic Heuristic Design (AHD), the authors introduce a novel Profile-Diverse Retention (PDR) strategy that retains diverse reformulation attempts rather than just the most recent or fastest. The findings reveal that iterative reformulation can yield significant speedups in solver performance, with diverse context retention strategies outperforming traditional methods.
Iterative reformulation using LLMs can dramatically accelerate solver efficiency by leveraging diverse context retention strategies.
Combinatorial problems appear in numerous industrial applications. A common approach is to formulate these problems as declarative constraint models that can subsequently be compiled to and solved by a range of back-end solvers. Recent work shows that Large Language Models (LLMs) can produce correct models from natural language, but even a correct model can be expensive to solve because performance remains sensitive to modelling choices. In this work, we investigate whether LLMs can automate performance-oriented model reformulation. Inspired by Automatic Heuristic Design (AHD), we use an evolutionary framework in which an LLM proposes candidate reformulations that are verified and benchmarked against the user-defined baseline model. We compare AHD-adapted search strategies that control which prior attempts, instructions, and measured feedback enter each prompt. Existing retention strategies prioritize recency or performance, but do not explicitly diversify the context. To cover this gap, we introduce Profile-Diverse Retention (PDR), which applies Maximal Marginal Relevance (MMR) to instance-level runtime vectors to retain behaviourally diverse attempts. We systematically evaluate the strategies on eight CSPLib problems using validation-based final model selection. The results show that: (i) iterative reformulation can produce substantial held-out speedups; (ii) strategies that keep the retained context diverse outperform those that retain only recent or the fastest attempts; and (iii) validation-based selection improves the held-out speedup of every strategy.
