AI Paper Atlas

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

Generalizable Heuristic Generation Through LLMs with Meta-Optimization

Yiding Shi, Jianan Zhou, Wen Song, Jieyi Bi, Yaoxin Wu, Zhiguang Cao, Jie Zhang

ICLR 2026 · Large language models › Agents, code and math · approach: LLMs for optimization and heuristics · 12 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 natural-language ideas to code generation (ablation)MoH-GLS · TSP1000optimality gap (%): 2.420% -> 1.710%Improved
add natural-language ideas to code generation (ablation)MoH-GLS · TSP500optimality gap (%): 1.644% -> 1.045%Improved
increase population size from 1 to 10 (ablation)MoH-GLS · TSP1000optimality gap (%): 3.300% -> 1.710%Improved
increase population size from 5 to 10 (ablation)MoH-GLS · TSP1000optimality gap (%): 2.338% -> 1.710%Improved
use o1-mini instead of GPT-4o-mini (ablation)MoH-GLS · TSP1000optimality gap (%): 1.710% -> 2.119%Hurt
use o1-mini instead of GPT-4o-mini (ablation)MoH-GLS · TSP100optimality gap (%): 0.035% -> 0.024%Mixed
use Deepseek-V3 instead of GPT-4o-mini (ablation)MoH-GLS · TSP1000optimality gap (%): 1.710% -> 2.527%Hurt
use Qwen-Plus instead of GPT-4o-mini (ablation)MoH-GLS · TSP1000optimality gap (%): 1.710% -> 3.277%Hurt
use MoH meta-optimizer instead of MCTS-AHD (constructive)MCTS-AHD · TSP average gapoptimality gap (%): 12.568% -> 11.792%Improved
use MoH-GLS instead of EoH-GLSEoH-GLS · TSP average gapoptimality gap (%): 0.721% -> 0.453%Improved
use MoH instead of EoH on online BPPEoH · Online BPP averageoptimality gap (%): 0.680% -> 0.453%Improved
use MoH-KGLS instead of MCTS-AHD-KGLSMCTS-AHD-KGLS · TSP average gapoptimality gap (%): 0.411% -> 0.391%Improved

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