Robust Heuristic Algorithm Design with LLMs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Karimi, Pantea, Rouhana, Dany, Namyar, Pooria, Kakarla, Siva Kesava Reddy, Arun, Venkat, Arzani, Behnaz
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914085631688704
author Karimi, Pantea
Rouhana, Dany
Namyar, Pooria
Kakarla, Siva Kesava Reddy
Arun, Venkat
Arzani, Behnaz
author_facet Karimi, Pantea
Rouhana, Dany
Namyar, Pooria
Kakarla, Siva Kesava Reddy
Arun, Venkat
Arzani, Behnaz
contents We posit that we can generate more robust and performant heuristics if we augment approaches using LLMs for heuristic design with tools that explain why heuristics underperform and suggestions about how to fix them. We find even simple ideas that (1) expose the LLM to instances where the heuristic underperforms; (2) explain why they occur; and (3) specialize design to regions in the input space, can produce more robust algorithms compared to existing techniques~ -- ~the heuristics we produce have a $\sim28\times$ better worst-case performance compared to FunSearch, improve average performance, and maintain the runtime.
format Preprint
id arxiv_https___arxiv_org_abs_2510_08755
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Robust Heuristic Algorithm Design with LLMs
Karimi, Pantea
Rouhana, Dany
Namyar, Pooria
Kakarla, Siva Kesava Reddy
Arun, Venkat
Arzani, Behnaz
Artificial Intelligence
Computation and Language
Networking and Internet Architecture
We posit that we can generate more robust and performant heuristics if we augment approaches using LLMs for heuristic design with tools that explain why heuristics underperform and suggestions about how to fix them. We find even simple ideas that (1) expose the LLM to instances where the heuristic underperforms; (2) explain why they occur; and (3) specialize design to regions in the input space, can produce more robust algorithms compared to existing techniques~ -- ~the heuristics we produce have a $\sim28\times$ better worst-case performance compared to FunSearch, improve average performance, and maintain the runtime.
title Robust Heuristic Algorithm Design with LLMs
topic Artificial Intelligence
Computation and Language
Networking and Internet Architecture
url https://arxiv.org/abs/2510.08755