Researchers have developed MAPLE, a novel agent designed to manage and update optimization problems through successive natural-language requests. This system integrates language-based problem construction with mathematical programming and evolutionary search, allowing it to retain previous optimization programs, accepted plans, and candidate solutions for future iterations. MAPLE was evaluated on the new NLDO benchmark, which includes 15 trajectories and 180 updates across various domains like scheduling and resource placement. The agent demonstrated strong performance, achieving an online scalar quality of 0.951 and a Pareto hypervolume ratio of 0.875, while also showing that maintaining executable state improves update validity and preserves search information. AI
IMPACT This research could make complex optimization problems more accessible to domain practitioners by enabling natural language interaction and dynamic adaptation.
RANK_REASON The item describes a new research paper detailing a novel agent for optimization problems. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
