Scribd
Upload
2 views37 pages

Lecture 9 - 22AIE201

Classical Planning involves generating a sequence of actions to achieve a goal in a deterministic and observable environment, with applications in fields like robotics and supply chain management. It utilizes the Planning Domain Definition Language (PDDL) to define actions and states, and can be approached through state-space or plan-space search algorithms. Key concepts include action schemas, state representations, and planning graphs, which help visualize and solve planning problems efficiently.

Uploaded by

vuppalashasank
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
2 views37 pages

Lecture 9 - 22AIE201

Classical Planning involves generating a sequence of actions to achieve a goal in a deterministic and observable environment, with applications in fields like robotics and supply chain management. It utilizes the Planning Domain Definition Language (PDDL) to define actions and states, and can be approached through state-space or plan-space search algorithms. Key concepts include action schemas, state representations, and planning graphs, which help visualize and solve planning problems efficiently.

Uploaded by

vuppalashasank
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 37

Lecture 9 - Unit 4

Classical Planning
Introduction

 Classical Planning : Process of generating a sequence


of actions to achieve a predefined goal within a
deterministic, discrete and fully observable
environment.
 Involves constructing a plan that transforms the initial
state of the environment into a desired goal state.
 Real life applications: Robotics, Supply chain
management.
Feature of Classical Planning

 Essential feature of classical reasoning is domain


independence.
 designed to apply to different problems without having
to learn domain knowledge or heuristics.
 enables the creation of general-purpose planners that
can solve problems.
 Main Objective: develop planning algorithms and
systems that can reason about and solve problems
Classical Planning

 How an agent can take advantage of the structure of a


problem to construct complex plan of actions?

 Devising a plan of action to achieve ones goal state.

 Factored representation - one in which a state of the


world is represented by a collection of variables.
Planning Domain Definition Language -
PDDL

 Allows to express all the actions required to solve a


problem into one action schema.

 Standard for application of the planning domains and


problems.

 Language of extension that helps to identify objects,


their properties and relationships between them.
Search Problem Definition -PDDL

 4 components:
 Initial State
 The actions that are available in a state
 The result of applying an action
 Goal State
State

 Represented as a conjunction of fluents that are ground,


functionless atoms.
a fluent is a term used to describe any condition or
property of the world that can change over time.
 E.g. Poor ∧ Unknown : represent the state of an
unfortunate agent.
A state in a package delivery might be At(Truck1,
Melbourne) ∧ At(Truck2, Sydney)
State: Representation

 A state can be treated either as a

 conjunction of fluents, which can be manipulated by


logical inference, or

 Set of fluents, which can be manipulated with set


operations.
State

 Representations not allowed in a state.

 At(x,y)

 ¬Poor

 At(Father(Fred),Sydney)
Actions

 Described by a set of action schemas that implicitly


define the ACTIONS(s) and RESULT(s,a) functions.

 Concentrates on problems where most actions leave


most things unchanged.

 PDDL specifies the result of an action in terms of what


changes; everything that stays the same is left
unmentioned.
Action Schema

 A set of ground of ground(variable-free) actions can


be represented by a single action schema.

 Action schema is a lifted representation – it lifts the


level of reasoning from propositional logic to a
restricted subset of FOL.
Action Schema

 Action name
 list of all variables used in the schema
 a precondition ; and
 an effect.
Action Schema

Precondition and Effect :Conjunction of literals (positive,


negative and atomic sentences).
Precondition : State in which an action can be executed.
Effect : The result of executing an action.
Action Schema

 An action a can be executed in state s if s entails the


precondition of a.
(a ∈ACTIONS(s)) ⇔ s ⊧ PRECOND(a),
Where any variables in a is universally quantified.

 ∀p,from,to (Fly(p,from,to) ∈ ACTIONS(s)) ⇔ s ⊧


(At(p,from) ∧Plane(p) ∧Airport(from) ∧Airport(to))
PDDL description of cargo transport
planning problem
Classical Planning Algorithms

 Classical planning algorithms can be broadly


categorized into two main approaches:
 State-Space Search
Make use of the state space by producing and concerning
the successive states recurrently until reaching a goal
state.
 Plan-Space Search
Applied in the operational plan space, where the
emphasis is on developing and fine-tuning the partial
maps.
State-Space Search Algorithms

 Forward (progression) state-space search


 Search through the possible states the agent can be in
by continuously applying actions from the initial state
to goal state.
 Planning process continuous as:
 Defining the Initial State
 Specifying Actions

 Exploring States
 Finding the Goal
How Forward Planning works?

 State Representations
 Actions and Transitions
 Search Algorithms
 BFS

 DFS

 A* and Greedy Best First Search


 Goal State
Advantages and Limitations of FSSP

 Deterministic and  Scalability


Predictable  Inefficiency in complex
 Systematic Exploration problems
 Easily implemented with  Limited to deterministic
classical algorithms environments.
Classical Planning Algorithms

 Backward (Regression) state-space search


 we start at the goal and apply the actions backward
until we find a sequence of steps that reaches the initial
state.
 also called relevant-state search.
State-Space Search Algorithms
Planning Graphs

 Visually representing possible states and actions that


aid in decision-making.
 Represent a planning problem as a series of levels –
states and actions.
 GraphPlan is an algorithm based on Planning Graph.
 Also used as a source of heuristics.
 Is an approximation of complete tree of all possible
actions and their results.
Blocks World Representation
Levels

 Action level and State level

 Initially : State level representing the initial state of the


planning problem.
State Levels

 Represents logical propositions or facts about the


world.

 Each successive state level contains all the propositions


of the previous level plus any that can be derived by
the actions of the intervening action levels.
Action Levels

 Represent actions.

 An action node connects to a state level if the state


contains all the preconditions necessary for that action.

 Actions in turn can create new state conditions,


influencing the subsequent state level.
Edges

 Two types of edges connecting:


 state nodes to action nodes
(indicating that the state meets the
preconditions for the action)
 action nodes to state nodes
(indicating the effects of the action)
Mutual Exclusion (Mutex) Relationships

 certain pairs of actions or states might be mutually


exclusive.

 critical for reducing the complexity of the planning


problem.
Levels in Planning Graph

• Level S0: initial state of the planning graph

• consists of nodes that represents the state or conditions


that can be true.

• Level A0: consists of nodes responsible for taking all


specific actions in terms of the S0.
• Si: state or condition which could hold at a time i, it
may be both P and ¬P.
• Ai: the actions that could have their preconditions
satisfied at i.
Planning Graph

 Make a plan graph of depth k.


 Search for a solution
 If succeed, return a plan
 Else, k=k+1
 Go to 1
Planning Graph

 Phase 1 : Planning Graph Construction


 Phase 2: Solution Extraction
CAKE Example
CAKE Example
CAKE Example
CAKE Example
CAKE Example
GRAPH PLAN ALGORITHM

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy