CS6260: Automated Planning and Learning (Jan-May 2026)

Course Information

Offered for: BTech, MTech, MS, PhD

Teaching Assistant: Sayak Sen, Swapnil S. Sonawane, and Tushar

Pre-requisite Skills: Familiarity with Logic, Programming, Analysis of Algorithms, and Probability

Grading Policy: Quiz 1: 20%
Quiz 2: 20%
End-semester: 30%
In class tasks: 5%
Assignments: 25%

Course Email: cs6260.jm26@gmail.com

Objectives

This course aims to introduce students to the theoretical foundations and practical techniques of automated planning. It covers a range of planning paradigms, from classical deterministic planning to advanced forms, including:

Temporal and durative actions
Planning under uncertainty
Hierarchical planning
Autonomously learning the underlying models for various planning paradigms

The overall goal is to enable AI systems to automatically generate sequences of high-level actions to achieve desired goals.

Course Contents

Module 1: Deterministic Planning [3 Weeks]

Introduction, Complexity, PDDL (Planning Domain Definition Language)
Planning with Deterministic Models
Heuristics, Planners, Planning Graphs
PDDL 3.0: Features and Planners
PDDL 3.1: Object Fluents and Practical Limitations to Adoption

Module 2: Hierarchical and Multi-Agent Planning [2 Weeks]

Hierarchical Task Networks (HTNs)
HDDL (Hierarchical Domain Definition Language) and HTN Planning
MA-PDDL (Multi-Agent PDDL) and Cooperative Multi-Agent Planning

Module 3: Planning under Uncertainty [3 Weeks]

Markov Decision Processes (MDPs) and RL Basics
FOND (Fully-Observable Non-Deterministic) Planning
Probabilistic Planning and PPDDL (Probabilistic PDDL)

Module 4: Learning for Planning [4 Weeks]

Learning Heuristics
Learning Deterministic Action Models
Learning HTN Methods
Learning PPDDL and NDRs (Noisy Deictic Rules)
Learning Predicates/Symbols for Planning

Module 5: Practical Applications (Some of the following topics depending on time) [2 Weeks]

PDDL for RL
PDDL for Task and Motion Planning (TAMP)
Large Language Models (LLMs) for PDDL Planning
Advanced applications of PDDL

Text Books

No single book covers all modules, so chapters from a couple of books will be referred to.

[APL] Acting, Planning, and Learning. Malik Ghallab, Dana Nau, and Paolo Traverso. Cambridge University Press, 2025. Freely available here.

[PDDL] An Introduction to the Planning Domain Definition Language. Patrik Haslum, Nir Lipovetzky, Daniele Magazzeni, and Christian Muise. Morgan & Claypool Publishers, 2019. Freely available through IITM here.

Other Reference Books

[MMAP] A Concise Introduction to Models and Methods for Automated Planning. Hector Geffner and Blai Bonet. Springer Nature, 2022. Freely available through IITM here.

[APA] Automated Planning and Acting. Malik Ghallab, Dana Nau, and Paolo Traverso. Cambridge University Press, 2016. Freely available here.

[APTP] Automated Planning: Theory and Practice. Malik Ghallab, Dana Nau, and Paolo Traverso. Morgan Kaufmann Publishers, 2004.

[AIMA] Artificial Intelligence: A Modern Approach, 3rd Edition. Stuart Russell and Peter Norvig. Prentice Hall, 2009. (Better for automated planning than the 4th edition)

Course Calendar

Tue, Jan 20

Introduction

APL Chapter 1