Guardado en:
Detalles Bibliográficos
Autores principales: Tang, Wenjie, Zhou, Yuan, Xu, Erqiang, Cheng, Keyan, Li, Minne, Xiao, Liquan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2503.06047
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918490123796480
author Tang, Wenjie
Zhou, Yuan
Xu, Erqiang
Cheng, Keyan
Li, Minne
Xiao, Liquan
author_facet Tang, Wenjie
Zhou, Yuan
Xu, Erqiang
Cheng, Keyan
Li, Minne
Xiao, Liquan
contents Large language model (LLM)-based agents are increasingly applied to complex strategic environments that demand long-horizon reasoning, multi-agent interaction, and decision-making under uncertainty. However, common existing benchmarks either assess isolated skills, lack environmental diversity, or rely on broad overall metrics. To address these issues, we introduce DSGBench, a more rigorous evaluation platform for strategic decision-making tasks. Firstly, it incorporates six complex strategic games which serve as ideal testbeds due to their long-term and multi-dimensional decision-making demands and flexibility in customizing tasks with various difficulty levels and targets. Secondly, DSGBench employs a fine-grained evaluation scoring system which examines the decision-making capabilities by looking into the performance in five specific dimensions, offering a comprehensive assessment in a better-designed fashion. Furthermore, DSGBench also incorporates an automated decision-tracking mechanism which enables in-depth analysis of agent behaviour patterns and the turning points in their strategies. We evaluate six popular LLM agents, including open-source and closed-source models, and observe distinct strengths and limitations among various tasks. Through decision trajectory analysis, we further identify systemic limitations in different LLMs. These findings offer valuable insights for model selection and future LLM-based agent development.
format Preprint
id arxiv_https___arxiv_org_abs_2503_06047
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DSGBench: A Diverse Strategic Game Benchmark for Evaluating LLM-based Agents in Complex Decision-Making Environments
Tang, Wenjie
Zhou, Yuan
Xu, Erqiang
Cheng, Keyan
Li, Minne
Xiao, Liquan
Artificial Intelligence
Computation and Language
Large language model (LLM)-based agents are increasingly applied to complex strategic environments that demand long-horizon reasoning, multi-agent interaction, and decision-making under uncertainty. However, common existing benchmarks either assess isolated skills, lack environmental diversity, or rely on broad overall metrics. To address these issues, we introduce DSGBench, a more rigorous evaluation platform for strategic decision-making tasks. Firstly, it incorporates six complex strategic games which serve as ideal testbeds due to their long-term and multi-dimensional decision-making demands and flexibility in customizing tasks with various difficulty levels and targets. Secondly, DSGBench employs a fine-grained evaluation scoring system which examines the decision-making capabilities by looking into the performance in five specific dimensions, offering a comprehensive assessment in a better-designed fashion. Furthermore, DSGBench also incorporates an automated decision-tracking mechanism which enables in-depth analysis of agent behaviour patterns and the turning points in their strategies. We evaluate six popular LLM agents, including open-source and closed-source models, and observe distinct strengths and limitations among various tasks. Through decision trajectory analysis, we further identify systemic limitations in different LLMs. These findings offer valuable insights for model selection and future LLM-based agent development.
title DSGBench: A Diverse Strategic Game Benchmark for Evaluating LLM-based Agents in Complex Decision-Making Environments
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2503.06047