Learn to Relax with Large Language Models: Solving Constraint Optimization Problems via Bidirectional Coevolution

1. 1Formulate the Constraint Optimization Problem (COP) to be solved
2. 2Implement the AutoCO method, which couples operations-research principles with Large Language Model (LLM) capabilities
3. 3Utilize bidirectional coevolution to enable the LLM to proactively design strategies
4. 4Evaluate and refine the solution generated by AutoCO

Researchers and engineers working on optimization problems and AI can benefit from this approach as it enables Large Language Models to proactively design strategies, rather than just passively checking constraints

💡 Large Language Models can be used as proactive strategy designers, rather than just passive constraint checkers, to solve complex optimization problems

💡 AutoCO: solving complex COPs with LLMs and bidirectional coevolution