AdaEvolve: Adaptive LLM Driven Zeroth-Order Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cemri, Mert, Agrawal, Shubham, Gupta, Akshat, Liu, Shu, Cheng, Audrey, Mang, Qiuyang, Naren, Ashwin, Erdogan, Lutfi Eren, Sen, Koushik, Zaharia, Matei, Dimakis, Alex, Stoica, Ion
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908848869081088
author Cemri, Mert
Agrawal, Shubham
Gupta, Akshat
Liu, Shu
Cheng, Audrey
Mang, Qiuyang
Naren, Ashwin
Erdogan, Lutfi Eren
Sen, Koushik
Zaharia, Matei
Dimakis, Alex
Stoica, Ion
author_facet Cemri, Mert
Agrawal, Shubham
Gupta, Akshat
Liu, Shu
Cheng, Audrey
Mang, Qiuyang
Naren, Ashwin
Erdogan, Lutfi Eren
Sen, Koushik
Zaharia, Matei
Dimakis, Alex
Stoica, Ion
contents The paradigm of automated program generation is shifting from one-shot generation to inference-time search, where Large Language Models (LLMs) function as semantic mutation operators within evolutionary loops. While effective, these systems are currently governed by static schedules that fail to account for the non-stationary dynamics of the search process. This rigidity results in substantial computational waste, as resources are indiscriminately allocated to stagnating populations while promising frontiers remain under-exploited. We introduce AdaEvolve, a framework that reformulates LLM-driven evolution as a hierarchical adaptive optimization problem. AdaEvolve uses an "accumulated improvement signal" to unify decisions across three levels: Local Adaptation, which dynamically modulates the exploration intensity within a population of solution candidates; Global Adaptation, which routes the global resource budget via bandit-based scheduling across different solution candidate populations; and Meta-Guidance which generates novel solution tactics based on the previously generated solutions and their corresponding improvements when the progress stalls. We demonstrate that AdaEvolve consistently outperforms the open-sourced baselines across 185 different open-ended optimization problems including combinatorial, systems optimization and algorithm design problems.
format Preprint
id arxiv_https___arxiv_org_abs_2602_20133
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AdaEvolve: Adaptive LLM Driven Zeroth-Order Optimization
Cemri, Mert
Agrawal, Shubham
Gupta, Akshat
Liu, Shu
Cheng, Audrey
Mang, Qiuyang
Naren, Ashwin
Erdogan, Lutfi Eren
Sen, Koushik
Zaharia, Matei
Dimakis, Alex
Stoica, Ion
Neural and Evolutionary Computing
Artificial Intelligence
Computation and Language
The paradigm of automated program generation is shifting from one-shot generation to inference-time search, where Large Language Models (LLMs) function as semantic mutation operators within evolutionary loops. While effective, these systems are currently governed by static schedules that fail to account for the non-stationary dynamics of the search process. This rigidity results in substantial computational waste, as resources are indiscriminately allocated to stagnating populations while promising frontiers remain under-exploited. We introduce AdaEvolve, a framework that reformulates LLM-driven evolution as a hierarchical adaptive optimization problem. AdaEvolve uses an "accumulated improvement signal" to unify decisions across three levels: Local Adaptation, which dynamically modulates the exploration intensity within a population of solution candidates; Global Adaptation, which routes the global resource budget via bandit-based scheduling across different solution candidate populations; and Meta-Guidance which generates novel solution tactics based on the previously generated solutions and their corresponding improvements when the progress stalls. We demonstrate that AdaEvolve consistently outperforms the open-sourced baselines across 185 different open-ended optimization problems including combinatorial, systems optimization and algorithm design problems.
title AdaEvolve: Adaptive LLM Driven Zeroth-Order Optimization
topic Neural and Evolutionary Computing
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2602.20133