BE - Artificial Intelligence

A Planning System is like a smart assistant that figures out how to achieve goals by breaking them down into step-by-step actions.

Main Components: 

1. Goal Description 

What you want to achieve 

    Defines the end state or objective
    Specifies what success looks like
    Example: "Make a sandwich" or "Win the chess game"
     

Key Features: 

    Clear and specific
    Measurable outcomes
    Time constraints (optional)
     

 
2. Current State Description 

Where you are now 

    Describes the present situation
    Lists all relevant facts and conditions
    Example: "I have bread, peanut butter, but no jelly"
     

Key Features: 

    Complete information about present
    Accurate representation of reality
    Updated as situation changes
     

 
3. Actions/Operators 

What you can do 

    Set of possible moves or steps
    Each action has preconditions and effects
    Example: "Spread peanut butter" requires having peanut butter and knife
     

Structure of each action: 

    Name: What the action is called
    Preconditions: What must be true before doing it
    Effects: What changes after doing it

4. State Space 

All possible situations 

    Collection of all reachable states
    Shows different paths and possibilities
    Like a map of all places you could go
     

Key Features: 

    Represents all valid configurations
    Shows connections between states
    Helps avoid impossible or duplicate states
     

 
5. Search Strategy 

How to find the best path 

    Method for exploring possible plans
    Decides which actions to try first
    Examples: Try everything, be smart about choices
     

Common Strategies: 

    Breadth-first: Try all short plans first
    Depth-first: Follow one path deeply
    Best-first: Choose most promising paths
    Heuristic: Use smart shortcuts
     

 
6. Plan Generator 

Creates the actual step-by-step plan 

    Combines all components into a sequence
    Orders actions logically
    Checks that plan makes sense
     

Process: 

    Start with current state
    Apply actions systematically
    Check if goal is reached
    If not, try different actions
    Output complete plan when found
     

 
7. Knowledge Base 

What the system knows 

    Facts about the world
    Rules about how things work
    Domain-specific information
    Example: "Fire needs oxygen" or "Bread goes stale"
     

Types of Knowledge: 

    Domain knowledge: Specific to the problem area
    Common sense: General world knowledge
    Procedural knowledge: How to do things
     

 
8. Execution Monitor 

Watches the plan in action 

    Tracks progress during plan execution
    Detects when things go wrong
    Triggers replanning when needed
     

Functions: 

    Compare actual vs. expected results
    Identify failures or changes
    Alert when plan needs adjustment
     

 
9. Plan Repair/Replanning Module 

Fixes problems when they occur 

    Modifies plans when reality differs
    Creates new plans when old ones fail
    Adapts to changing conditions
     

When activated: 

    Action fails unexpectedly
    New information discovered
    Goals change during execution