Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 1 Review for the Final Exam Dana S. Nau Dept. of Computer Science, and Institute for Systems Research University of Maryland Lecture slides for Automated Planning: Theory and Practice
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 2 The Chapters We’ve Covered l Chapter 1: Introduction l Chapter 2: Representations for Classical Planning l Chapter 3: Complexity of Classical Planning l Chapter 4: State-Space Planning l Chapter 5: Plan-Space Planning l Chapter 6: Planning-Graph Techniques l Chapter 7: Propositional Satisfiability Techniques l Chapter 10: Control Rules in Planning l Chapter 11: Hierarchical Task Network Planning l Chapter 23: Planning in the Game of Bridge l Chapter 16: Planning based on Markov Decision Processes l Chapter 17: Planning based on Model Checking l Chapter 9: Heuristics in Planning l Chapter 14: Temporal Planning l Chapter 15: Planning and Resource Scheduling l Chapter 21: Planning for Manufacturability Analysis
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 3 Chapter 1: Introduction and Overview l 1.1: First Intuitions on Planning l 1.2: Forms of planning l 1.3: Domain-Independent Planning l 1.4: Conceptual Model for Planning l 1.5: Restricted Model l 1.6: Extended Models l 1.7: A Running Example: Dock-Worker Robots Not very much I could ask about
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 4 2: Representations for Classical Planning l 2.1: Introduction l 2.2: Set-Theoretic Representation u 2.2.1: Planning Domains, Problems, and Solutions u 2.2.2: State Reachability u 2.2.3: Stating a Planning Problem u 2.2.4: Properties of the Set-theoretic Representation l 2.3: Classical Representation u 2.3.1: States u 2.3.2: Operators and Actions u 2.3.3: Plans, Problems, and Solutions u 2.3.4: Semantics of Classical Representations
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 5 2: Representations for Classical Planning l 2.4: Extending the Classical Representation u 2.4.1: Simple Syntactical Extensions u 2.4.2: Conditional Planning Operators u 2.4.3: Quantified Expressions u 2.4.4: Disjunctive Preconditions u 2.4.5: Axiomatic Inference u 2.4.6: Function Symbols u 2.4.7: Attached Procedures u 2.4.8: Extended Goals l 2.5: State-Variable Representation u 2.5.1: State Variables u 2.5.2: Operators and Actions u 2.5.3: Domains and Problems u 2.5.4: Properties l 2.6: Comparisons See Chapters 10, 11, …
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 6 2: Representations for Classical Planning l 2.5: State-Variable Representation u 2.5.1: State Variables u 2.5.2: Operators and Actions u 2.5.3: Domains and Problems u 2.5.4: Properties l 2.6: Comparisons
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 7 Chapter 3: Complexity of Classical Planning l 3.1: Introduction l 3.2: Preliminaries l 3.3: Decidability and Undecidability Results l 3.4: Complexity Results u 3.4.1: Binary Counters u 3.4.2: Unrestricted Classical Planning u 3.4.3: Other results l 3.5: Limitations Not very much I could ask about
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 8 Chapter 4: State-Space Planning l 4.1: Introduction l 4.2: Forward Search u 4.2.1: Formal Properties u 4.2.2: Deterministic Implementations l 4.3: Backward Search l 4.4: The STRIPS Algorithm l 4.5: Domain-Specific State-Space Planning u 4.5.1: The Container-Stacking Domain u 4.5.2: Planning Algorithm soundness & completeness, loop checking, relevant actions See Chapters 10 and 11
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 9 Chapter 5: Plan-Space Planning l 5.1: Introduction l 5.2: The Search Space of Partial Plans l 5.3: Solution Plans l 5.4: Algorithms for Plan Space Planning u 5.4.1: The PSP Procedure u 5.4.2: The PoP Procedure l 5.5: Extensions l 5.6: Plan Space Versus State Space Planning flaws, protected conditions, resolvers
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 10 Chapter 6: Planning-Graph Techniques l 6.1: Introduction l 6.2: Planning Graphs u 6.2.1: Reachability Trees u 6.2.2: Reachability with Planning Graphs u 6.2.3: Independent Actions and Layered Plans u 6.2.4: Mutual Exclusion Relations l 6.3: The Graphplan Planner u 6.3.1: Expanding the Planning Graph u 6.3.2: Searching the Planning Graph u 6.3.3: Analysis of Graphplan l 6.4: Extensions and Improvements of Graphplan u 6.4.1: Extending the Language u 6.4.2: Improving the Planner u 6.4.3: Extending the Independence Relation pay more attention to my notes than to the chapter
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: : Propositional Satisfiability Techniques l 7.1: Introduction l 7.2: Planning problems as Satisfiability problems u 7.2.1: States as propositional formulas u 7.2.2: State transitions as propositional formulas u 7.2.3: Planning problems as propositional formulas l 7.3: Planning by Satisfiability u 7.3.1: Davis-Putnam u 7.3.2: Stochastic Procedures l 7.4: Different Encodings u 7.4.1: Action Representation u 7.4.2: Frame axioms I probably won’t ask you about these procedures. If I change my mind, I’ll let you know.
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 12 Chapter 10: Control Rules in Planning l Intro to Part III: Heuristics and Control Strategies l 10.1: Introduction l 10.2: Simple Temporal Logic l 10.3: Progression l 10.4: Planning Procedure l 10.5: Extensions l 10.6: Extended Goals
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 13 Chapter 11: HTN Planning l 11.1: Introduction l 11.2: STN Planning u : Tasks and Methods u : Problems and Solutions l 11.3: Total-Order STN Planning l 11.4: Partial-Order STN Planning l 11.5: HTN Planning u : Task Networks u : HTN Methods u : HTN Problems and Solutions u : Planning Procedures
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 14 Chapter 11: HTN Planning l 11.6: Comparisons u : HTN Planning Versus STN Planning u : HTN Methods Versus Control Rules l 11.7: Extensions u : Extensions from Chapter 2 u : Additional Extensions l 11.8: Extended Goals
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 15 Chapter 16: Planning based on MDPs l 16.1: Introduction l 16.2: Planning in Fully Observable Domains u : Domains, Plans, and Planning Problems u : Planning Algorithms l 16.3: Planning under Partial Observability u : Domains, Plans, and Planning Problems u : Planning Algorithms l 16.4: Reachability and Extended Goals
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: : Planning based on Model Checking l 17.1: Introduction l 17.2: Planning for Reachability Goals u : Domains, Plans, and Planning Problems u : Planning Algorithms l 17.3: Planning for Extended Goals u : Domains, Plans, and Planning Problems u : Planning Algorithms u : Beyond Temporal Logics l 17.4: Planning under Partial Observability u : Domains, Plans, and Planning Problems u : Planning Algorithms l 17.5: Planning as Model Checking vs. MDPs
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 17 Chapter 9: Heuristics in Planning l 9.1: Introduction l 9.2: Design Principle for Heuristics: Relaxation l 9.3: Heuristics for State-Space Planning u 9.3.1: State Reachability Relaxation u 9.3.2: Heuristically Guided Backward Search u 9.3.3: Admissible State-Space Heuristics u 9.3.4: Graphplan as a Heuristic-Search Planner l 9.4: Heuristics for Plan-Space Planning u 9.4.1: Flaw-Selection Heuristics u 9.4.2: Resolver-Selection Heuristics
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 18 Chapter 14: Temporal Planning l 14.1: Introduction l 14.2: Planning with Temporal Operators u : Temporal Expressions and Temporal Databases u : Temporal Planning Operators u : Domain axioms u : Temporal Planning Domains, Problems and Plans u : Concurrent Actions with Interfering Effects u : A Temporal Planning Procedure l 14.3: Planning with Chronicles u : State Variables, Timelines and Chronicles u : Chronicles as Planning Operators u : Chronicle Planning Procedures u : Constraint Management in CP u : Search Control in CP
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: : Planning and Resource Scheduling l 15.1: Introduction l 15.2: Elements of Scheduling Problems u : Actions u : Resources u : Constraints and Cost Functions l 15.3: Machine Scheduling Problems u : Classes of Machine Scheduling Problems u : Complexity of Machine Scheduling u : Solving Machine Scheduling Problems u : Planning and Machine Scheduling l 15.4: Integrating Planning and Scheduling u : Representation u : Detecting Resource Conflicts u : Managing Resource-Conflict Flaws
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 20 The Exam l December 15, CSI 1121, 1:30pm to 3:30pm l Open book, open notes (closed neighbor) l On the exam sheet, I’ll let you choose u Midterm 20%, Final 50% u Midterm 30%, Final 40% u Midterm 40%, Final 30% l The other 30% is the project
Dana Nau: Lecture slides for Automated Planning Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License: 21 Miscellaneous l If you have questions: u l Sample exams: go to the home page, click on “sample exams” u Send if you’ve forgotten the name/password u Midterm and final for four different semesters »The more recent the semester, the closer the material is to what we covered this time