Introduction

• agent:
• state: (s) agent in its environment
• initial state: starting point
• actions: (a) -> bring from state to state
• transition model:
• state space:
• -> graph
• goal state: end state(s)
• path cost: cost associated with given path

Search problem

• initial state
• actions
• transition model
• goal test
• path cost function

Node

data structure that keep track of

• a state
• a parent (node generated this node)
• an action (applied to parent to get node)
• a path cost (from initia;l state to node)

Approach

• start with a frontier that contains the initial state
• start with an empty explored set
• repeat
  • frontier == empty -> no solution
  • remove a node from the frontier
  • if node == goal state -> return solution
  • add the node to the explored set
  • else
    • expand node add resulting nodes to the frontier if they arent in the frontier or the explored set

Stack (used for frontier)

• last in first out (LIFO)

DEPTH-FIRST-SEARCH

Breath-First-Search

• explore the shallowest node in the frontier
• first in first out (FIFO)

knowledge

learning

• Supervised Learning
• Reinforcement Learning
• Unsupervised Learning