
CC
Camilo Chac\'on Sartori, Guillem Rodr\'iguez-Corominas, Christian Blum
· 1 min read
ResearcharXiv cs.CL
From Hand-Crafted to LLM-Based Variation Operators in Metaheuristics: A Tutorial
arXiv:2609.31649v1 Announce Type: cross
Abstract: Large language models (LLMs) are increasingly being employed as variation operators in metaheuristics, generating or modifying candidate solutions, heuristics, or programs inside iterative search loops. This shift reframes variation as a model call conditioned on different types of information. We introduce an operator-level framework with two descriptors: (1) the type of prompt-conditioning information at variation time (\texttt{Numeric}, \texttt{Symbolic}, \texttt{Linguistic}), and (2) artifact persistence, identifying what survives the model call (\texttt{Transient}, \texttt{Amortized}, \texttt{Transfer}). The tutorial shows how to classify, build, and select these operators through a worked build template, a method survey, an evidence table, and a cost-aware decision guide.
Original source
This story was published by arXiv cs.CL and written by Camilo Chac\'on Sartori, Guillem Rodr\'iguez-Corominas, Christian Blum. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on arxiv.org


