AI Paper Atlas

Research map / Large language models / Agents, code and math

Pareto-Grid-Guided Large Language Models for Fast and High-Quality Heuristics Design in Multi-Objective Combinatorial Optimization

Ha Minh Hieu, Hung Phan, Tung Duy Doan, Tung Dao, Cong Dao Tran, Huynh Thi Thanh Binh

AAAI 2026 · Large language models › Agents, code and math · approach: LLMs for optimization and heuristics · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add LLM reflection feedback (ablation)MPaGE · Bi-TSP20HV: 0.829 -> 0.941Improved
use LLM semantic clustering instead of AST similarity (ablation)MPaGE · Bi-TSP20HV: 0.745 -> 0.921Improved
use LLM semantic clustering instead of K-Means (ablation)MPaGE · Bi-TSP20HV: 0.876 -> 0.921Improved
use LLM semantic clustering instead of SWDI clustering (ablation)MPaGE · Tri-TSP20HV: 0.888 -> 0.892No clear effect
use Pareto Front Grid backbone instead of NSGA-II (ablation)MPaGE · Bi-TSP100HV: 0.757 -> 0.844Improved
use Pareto Front Grid backbone instead of NSGA-II (ablation)MPaGE · Bi-TSP100IGD: 0.095 -> 0.099Hurt
use MPaGE instead of MEoHLLM heuristic design · Bi-TSP20HV: 0.724 -> 0.911Improved
use MPaGE instead of ReEvoLLM heuristic design · Bi-KP50HV: 0.996 -> 0.932Hurt
use MPaGE heuristics instead of PFG-MOEABi-KP200 · Bi-KP200HV: 0.330 -> 0.197Hurt
use MPaGE heuristics instead of NSGA-IIBi-CVRP100 · Bi-CVRP100HV: 0.357 -> 0.422Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy