# 面向大语言模型智能体的环境工程综述

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-06-10 08:00
- AIHOT 分数：69
- AIHOT 链接：https://aihot.virxact.com/items/cmq8ws2lj06ihslldlx4bemi4
- 原文链接：https://arxiv.org/abs/2606.12191

## AI 摘要

该综述系统研究了大语言模型智能体的环境工程生命周期，涵盖环境建模、合成、评估与应用。从八个属性和八个领域梳理代表性环境的发展路径；归纳自动环境合成的符号合成与神经合成两种范式及对应评估方法。从智能体-环境共同演化视角，总结四种智能体演化路径（记忆中心、编排中心、轨迹中心、探索中心）与三种环境演化范式（神经驱动、难度驱动、规模驱动）。最后展望环境即服务（EaaS）、多智能体环境和神经符号环境等方向。

## 正文

Environments serve as interactive systems for large language model (LLM) based agents across diverse scenarios and play a crucial role in driving the continual evolution of model capabilities. Despite this importance, existing work lacks a systematic categorization and deep analysis. This paper systematically studies current researches on agentic environments from the perspective of the environment engineering lifecycle, covering their modeling, synthesis, evaluation and application. Specifically, the paper first introduces representative environments from the perspectives of eight attributes and eight domains, providing detailed analyses of their development paths and highlighting their core capabilities. Second, for automated environment synthesis, two paradigms are introduced, such as symbolic synthesis and neural synthesis. This paper also shows different environment evaluation methods in each paradigm. Thirdly, the corresponding environment applications from the perspective of agent-environment co-evolution are discussed. In specific, the paper characterizes the primary pathways for agent evolution in dynamic environments from four complementary perspectives: memory-centric experience evolution, orchestration-centric workflow evolution, trajectory-centric offline evolution, and exploration-centric online evolution. And three paradigms of environment evolution are identified, namely neural-driven, difficulty-driven, and scaling-driven approaches. At last, several promising future directions are discussed, including Environment-as-a-Service, Multi-agent Environments, and Neural-Symbolic Environments.