LLM-based Agent Planning Module

An LLM-based Agent Planning Module is a LLM-based system module that enables planning and reasoning.

• Context:
  • It can (typically) be referenced in an LLM-based Agent Architecture.
  • It can allow breaking down tasks into subgoals.
  • It can generate plans with or without external feedback.
  • It can aid in multi-step decision making.
  • It can employ techniques like chain-of-thought prompting.
  • ...
• Example(s):
  • The planning module in (Wang et al., 2023).
  • The goal decomposition planner in Voyager AI Agent (Wang, Xie et al., 2023).
  • The reasoning module in AutoGPT (Hou et al., 2023).
  • ...
• Counter-Example(s):
  • an End-to-End LLM systems with no separate planning component.
• See: LLM Agent Architecture, AI Planning.

---

---

References

2023

• (Wang, Ma et al., 2023) ⇒ Lei Wang, Chen Ma, Xueyang Feng, ..., and Ji-Rong Wen. (2023). “A Survey on Large Language Model based Autonomous Agents.” In: arXiv preprint arXiv:2308.11432. doi:10.48550/arXiv.2308.11432
  • QUOTE: ...
  • SUMMARY: It presents a unified framework for LLM agent architecture with 4 modules: LLM-based agent profile module, LLM-based agent memory module, LLM-based agent planning module, LLM-based agent action module.
  • 
    Figure 2: A unified framework for the architecture design of LLM-based autonomous AI agent.

---