Verified80% confidenceFactTime unknown

Plan-and-Execute和ReAct是当前Agent领域两种代表性的规划范式

10

Sources

80%

Confidence

Long-term

Relevance

6/1/2026

First Seen

Sources

GIS-agent：自然语言驱动ArcGIS的开源地理空间AI智能体

githubCalamityXeroc

Related Entities

Plan-and-Executeconcept ReActconcept

Related Claims

VerifiedAgent的实现通常通过ReAct（Reasoning + Acting）框架来实现规划能力78% similar Unverified自主代理的规划-执行-反馈循环架构最早在ReAct（Reasoning + Acting）论文中被系统化描述77% similar VerifiedReAct（Reasoning + Acting）架构中Agent在每一步先推理当前状态再决定下一步行动，形成感知-规划-执行的闭环75% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/28526

API

curl https://kongchang.com/api/v1/knowledge/claims/28526

MCP

get_claim(id=28526)