Subtopic Deep Dive
Case-Based Reasoning Planning
Research Guide
What is Case-Based Reasoning Planning?
Case-Based Reasoning Planning integrates CBR cycles of retrieve, reuse, revise, and retain with HTN and partial-order planning techniques to reuse past experiences for solving novel planning problems.
This approach leverages adaptation knowledge and competency models to improve planning robustness (Nau et al., 2003). It combines case retrieval from knowledge bases like CYC with hierarchical task network planning in SHOP2 (Lenat, 1995; Nau et al., 2003). Over 900 citations document SHOP2's role in international planning competitions.
Why It Matters
Case-Based Reasoning Planning enhances robustness in dynamic environments like space missions by adapting prior plans, as enabled by PDDL extensions for temporal domains (Fox and Long, 2003). Systems like Watson used similar knowledge integration for real-time decision-making under uncertainty (Ferrucci et al., 2010). SHOP2's HTN methods support efficient planning reuse, outperforming in competitions (Nau et al., 2003). Decision-theoretic models further leverage structural assumptions for scalable planning (Boutilier et al., 1999).
Key Research Challenges
Case Adaptation Scalability
Adapting retrieved cases to novel situations scales poorly with large case bases, requiring efficient revision mechanisms (Nau et al., 2003). HTN integration demands competency models to predict adaptation success (Fox and Long, 2003). Knowledge-rich systems like CYC highlight encoding challenges for reuse (Lenat, 1995).
Uncertainty in Retrieval
Retrieving relevant cases under uncertainty demands decision-theoretic planning to resolve ambiguities (Boutilier et al., 1999). Blackboard architectures like Hearsay-II integrate multi-level knowledge but struggle with real-time planning (Erman et al., 1980). Temporal extensions complicate similarity matching (Fox and Long, 2003).
Competency Model Learning
Building accurate competency models for CBR-HTN planners requires explanation-based learning to generalize cases (DeJong and Mooney, 1986). Portfolio selection aids solver choice but needs CBR adaptation (Xu et al., 2008). Causal representations could improve but remain underexplored (Schölkopf et al., 2021).
Essential Papers
CYC
Douglas B. Lenat · 1995 · Communications of the ACM · 1.9K citations
Since 1984, a person-century of effort has gone into building CYC, a universal schema of roughly 10 5 general concepts spanning human reality. Most of the time has been spent codifying knowledge ab...
PDDL2.1: An Extension to PDDL for Expressing Temporal Planning Domains
Maria Fox, Derek Long · 2003 · Journal of Artificial Intelligence Research · 1.7K citations
In recent years research in the planning community has moved increasingly toward s application of planners to realistic problems involving both time and many typ es of resources. For example, inter...
Building Watson: An Overview of the DeepQA Project
David Ferrucci, Eric W. Brown, Jennifer Chu‐Carroll et al. · 2010 · AI Magazine · 1.5K citations
IBM Research undertook a challenge to build a computer system that could compete at the human champion level in real time on the American TV quiz show, Jeopardy. The extent of the challenge include...
The Hearsay-II Speech-Understanding System: Integrating Knowledge to Resolve Uncertainty
Lee D. Erman, Frederick Hayes‐Roth, Victor Lesser et al. · 1980 · ACM Computing Surveys · 1.3K citations
article Free Access Share on The Hearsay-II Speech-Understanding System: Integrating Knowledge to Resolve Uncertainty Authors: Lee D. Erman USC/Information Sciences Institute, Marina del Rey, Calif...
Decision-Theoretic Planning: Structural Assumptions and Computational Leverage
Craig Boutilier, Taraneh Dean, Steve Hanks · 1999 · Journal of Artificial Intelligence Research · 1.1K citations
Planning under uncertainty is a central problem in the study of automated sequential decision making, and has been addressed by researchers in many different fields, including AI planning, decision...
SHOP2: An HTN Planning System
Dana Nau, Tsz-Chiu Au, Okhtay Ilghami et al. · 2003 · Journal of Artificial Intelligence Research · 922 citations
The SHOP2 planning system received one of the awards for distinguished performance in the 2002 International Planning Competition. This paper describes the features of SHOP2 which enabled it to exc...
Toward Causal Representation Learning
Bernhard Schölkopf, Francesco Locatello, Stefan Bauer et al. · 2021 · Proceedings of the IEEE · 899 citations
The two fields of machine learning and graphical causality arose and are developed separately. However, there is, now, cross-pollination and increasing interest in both fields to benefit from the a...
Reading Guide
Foundational Papers
Start with SHOP2 (Nau et al., 2003) for HTN-CBR implementation, PDDL2.1 (Fox and Long, 2003) for temporal domains, and CYC (Lenat, 1995) for knowledge-rich case bases.
Recent Advances
Study decision-theoretic leverage (Boutilier et al., 1999) and causal representations (Schölkopf et al., 2021) for modern uncertainty handling.
Core Methods
Core techniques: CBR 4R-cycle, SHOP2 HTN decomposition, PDDL temporal modeling, blackboard integration from Hearsay-II (Erman et al., 1980).
How PapersFlow Helps You Research Case-Based Reasoning Planning
Discover & Search
Research Agent uses searchPapers and citationGraph to map CBR-HTN evolution from SHOP2 (Nau et al., 2003) to PDDL2.1 (Fox and Long, 2003), revealing 922+ citations linking to decision-theoretic works. exaSearch uncovers adaptation-focused papers beyond top lists, while findSimilarPapers expands from CYC (Lenat, 1995) to knowledge representation handbooks.
Analyze & Verify
Analysis Agent employs readPaperContent on SHOP2 to extract HTN-CBR algorithms, then verifyResponse with CoVe checks claims against PDDL2.1 temporal models (Fox and Long, 2003). runPythonAnalysis simulates case retrieval similarity metrics using NumPy/pandas on extracted data, with GRADE scoring evidence strength for competency models.
Synthesize & Write
Synthesis Agent detects gaps in case adaptation post-SHOP2 via contradiction flagging across Boutilier et al. (1999) and Nau et al. (2003). Writing Agent applies latexEditText and latexSyncCitations to draft HTN-CBR surveys, using latexCompile for publication-ready output and exportMermaid for CBR cycle diagrams.
Use Cases
"Implement SHOP2 case retrieval in Python for HTN planning benchmark."
Research Agent → searchPapers('SHOP2 HTN') → Code Discovery → paperExtractUrls → paperFindGithubRepo → runPythonAnalysis → benchmarked retrieval code with NumPy similarity metrics.
"Compare CBR adaptation in SHOP2 vs PDDL2.1 temporal planning."
Research Agent → citationGraph → Analysis Agent → readPaperContent (Nau et al., 2003; Fox and Long, 2003) → Synthesis → latexEditText → latexSyncCitations → latexCompile → formatted LaTeX comparison table.
"Find GitHub repos adapting CYC knowledge for modern CBR planners."
Research Agent → exaSearch('CYC case-based planning') → Code Discovery → paperFindGithubRepo('Lenat 1995') → githubRepoInspect → verified adaptation code examples.
Automated Workflows
Deep Research workflow scans 50+ papers from Nau et al. (2003) citation graph, producing structured CBR-HTN reviews with GRADE-verified claims. DeepScan's 7-step chain analyzes SHOP2 (Nau et al., 2003) via readPaperContent → CoVe → runPythonAnalysis for temporal competency simulations. Theorizer generates novel adaptation theories from CYC-SHOP2 patterns (Lenat, 1995; Nau et al., 2003).
Frequently Asked Questions
What defines Case-Based Reasoning Planning?
It combines CBR cycles (retrieve, reuse, revise, retain) with HTN planning like SHOP2 for experience reuse (Nau et al., 2003).
What are core methods in CBR Planning?
Methods include case retrieval via similarity metrics, HTN decomposition in SHOP2, and temporal extensions from PDDL2.1 (Nau et al., 2003; Fox and Long, 2003).
What are key papers on CBR Planning?
Seminal works are SHOP2 (Nau et al., 2003, 922 citations), PDDL2.1 (Fox and Long, 2003, 1703 citations), and CYC knowledge base (Lenat, 1995, 1936 citations).
What open problems exist in CBR Planning?
Challenges include scalable adaptation under uncertainty and learning competency models, linking decision-theoretic planning (Boutilier et al., 1999) with causal learning (Schölkopf et al., 2021).
Research AI-based Problem Solving and Planning with AI
PapersFlow provides specialized AI tools for Computer Science researchers. Here are the most relevant for this topic:
AI Literature Review
Automate paper discovery and synthesis across 474M+ papers
Code & Data Discovery
Find datasets, code repositories, and computational tools
Deep Research Reports
Multi-source evidence synthesis with counter-evidence
AI Academic Writing
Write research papers with AI assistance and LaTeX support
See how researchers in Computer Science & AI use PapersFlow
Field-specific workflows, example queries, and use cases.
Start Researching Case-Based Reasoning Planning with AI
Search 474M+ papers, run AI-powered literature reviews, and write with integrated citations — all in one workspace.
See how PapersFlow works for Computer Science researchers