GA4GC: Greener Agent for Greener Code via Multi-Objective Configuration Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gong, Jingzhi, Bian, Yixin, de la Cal, Luis, Pinna, Giovanni, Uteem, Anisha, Williams, David, Zamorano, Mar, Even-Mendoza, Karine, Langdon, W. B., Menendez, Hector, Sarro, Federica
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915533766524928
author Gong, Jingzhi
Bian, Yixin
de la Cal, Luis
Pinna, Giovanni
Uteem, Anisha
Williams, David
Zamorano, Mar
Even-Mendoza, Karine
Langdon, W. B.
Menendez, Hector
Sarro, Federica
author_facet Gong, Jingzhi
Bian, Yixin
de la Cal, Luis
Pinna, Giovanni
Uteem, Anisha
Williams, David
Zamorano, Mar
Even-Mendoza, Karine
Langdon, W. B.
Menendez, Hector
Sarro, Federica
contents Coding agents powered by LLMs face critical sustainability and scalability challenges in industrial deployment, with single runs consuming over 100k tokens and incurring environmental costs that may exceed optimization benefits. This paper introduces GA4GC, the first framework to systematically optimize coding agent runtime (greener agent) and code performance (greener code) trade-offs by discovering Pareto-optimal agent hyperparameters and prompt templates. Evaluation on the SWE-Perf benchmark demonstrates up to 135x hypervolume improvement, reducing agent runtime by 37.7% while improving correctness. Our findings establish temperature as the most critical hyperparameter, and provide actionable strategies to balance agent sustainability with code optimization effectiveness in industrial deployment.
format Preprint
id arxiv_https___arxiv_org_abs_2510_04135
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GA4GC: Greener Agent for Greener Code via Multi-Objective Configuration Optimization
Gong, Jingzhi
Bian, Yixin
de la Cal, Luis
Pinna, Giovanni
Uteem, Anisha
Williams, David
Zamorano, Mar
Even-Mendoza, Karine
Langdon, W. B.
Menendez, Hector
Sarro, Federica
Software Engineering
Artificial Intelligence
Coding agents powered by LLMs face critical sustainability and scalability challenges in industrial deployment, with single runs consuming over 100k tokens and incurring environmental costs that may exceed optimization benefits. This paper introduces GA4GC, the first framework to systematically optimize coding agent runtime (greener agent) and code performance (greener code) trade-offs by discovering Pareto-optimal agent hyperparameters and prompt templates. Evaluation on the SWE-Perf benchmark demonstrates up to 135x hypervolume improvement, reducing agent runtime by 37.7% while improving correctness. Our findings establish temperature as the most critical hyperparameter, and provide actionable strategies to balance agent sustainability with code optimization effectiveness in industrial deployment.
title GA4GC: Greener Agent for Greener Code via Multi-Objective Configuration Optimization
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2510.04135