ACL Findings2026
What Makes an LLM a Good Optimizer? A Trajectory Analysis of LLM-Guided Evolutionary Search
Xinhao Zhang, Xi Chen, François Portet, Maxime Peyrard
Abstract
Recent work has demonstrated the promise of orchestrating large language models (LLMs) within evolutionary and agentic optimization systems. However, the mechanisms driving these optimization gains remain poorly understood. In this work, we present a large-scale study of LLM-guided evolutionary search, collecting optimization trajectories for 15 LLMs across 8 tasks. Although zero-shot problem-solving ability correlates with final optimization outcomes, it explains only part of the variance: models with similar initial capability often induce dramatically different search trajectories and outcomes. By analyzing these trajectories, we find that strong LLM optimizers behave as local refiners, producing frequent incremental improvements while progressively localizing the search in semantic space. Conversely, weaker optimizers exhibit large semantic drift, with sporadic breakthroughs followed by stagnation. Notably, various measures of solution novelty do not predict final performance; novelty is beneficial only when the search remains sufficiently localized around high-performing regions of the solution space. Our results highlight the importance of trajectory analysis for understanding and improving LLM-based optimization systems and provide actionable insights for their design and training.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- Promptbreeder: Self-Referential Self-Improvement via Prompt Evolution cites Large Language Models as Optimizers
- Large Language Models as Optimizers cites Promptbreeder: Self-Referential Self-Improvement via Prompt Evolution
- Promptbreeder: Self-Referential Self-Improvement via Prompt Evolution cites Connecting Large Language Models with Evolutionary Algorithms Yields Powerful Prompt Optimizers
- ReEvo: Large Language Models as Hyper-Heuristics with Reflective Evolution cites Large Language Models as Optimizers
- ReEvo: Large Language Models as Hyper-Heuristics with Reflective Evolution cites Connecting Large Language Models with Evolutionary Algorithms Yields Powerful Prompt Optimizers
- Large Language Models as Optimizers cites Connecting Large Language Models with Evolutionary Algorithms Yields Powerful Prompt Optimizers
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning cites Promptbreeder: Self-Referential Self-Improvement via Prompt Evolution
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning cites Large Language Models as Optimizers
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning cites Connecting Large Language Models with Evolutionary Algorithms Yields Powerful Prompt Optimizers
Builds on13
- Large Language Models as OptimizersChengrun Yang, Xuezhi Wang, Yifeng Lu, Hanxiao Liu et al.ICLR 2024 · 817 citations
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement LearningLakshya A. Agrawal, Shangyin Tan, Dilara Soylu, Noah Ziems et al.ICLR 2026 · 466 citations
- Promptbreeder: Self-Referential Self-Improvement via Prompt EvolutionChrisantha Fernando, Dylan Banarse, Henryk Michalewski, Simon Osindero et al.ICML 2024 · 432 citations
- ReEvo: Large Language Models as Hyper-Heuristics with Reflective EvolutionHaoran Ye, Jiarui Wang, Zhiguang Cao, Federico Berto et al.NeurIPS 2024 · 424 citations
- Connecting Large Language Models with Evolutionary Algorithms Yields Powerful Prompt OptimizersQingyan Guo, Rui Wang, Junliang Guo, Bei Li et al.ICLR 2024 · 257 citations
Related papers
- LASeR: Towards Diversified and Generalizable Robot Design with Large Language ModelsJunru Song, Yang Yang, Huan Xiao, Wei Peng et al.ICLR 2025
- Advancing Reasoning with Off-the-Shelf LLMs: A Semantic Structure PerspectivePengfei He, Zitao Li, Yue Xing, Yaliang Li et al.EMNLP Findings 2025
- Experience-Driven Reflective Co-Evolution of Prompts and Heuristics for Autonomous Algorithm DesignYihong Liu, Junyi Li, Hongyu Lu, Xin Zhao et al.ACL Findings 2026
- SwarmAgentic: Towards Fully Automated Agentic System Generation via Swarm IntelligenceYao Zhang, Chenyang Lin, Shijie Tang, Haokun Chen et al.EMNLP 2025 · 2 citations
- HiFo-Prompt: Prompting with Hindsight and Foresight for LLM-based Automatic Heuristic DesignChentongChen, Mengyuan Zhong, Jialong Shi, Jianyong Sun et al.ICLR 2026 · 17 citations
