Rational agent

agent: senses the environment and acts upon the environment

rational agent selects actions that maximize its (expected) utility

Agent

State: something that the environment can be

State space: all possible state

Actions: finite set of executable actions from a given state to another. This forms a path

Successors: the state that results from taking a given action to a given state

Cost: resources related to a given action

Initial state: the state to start the search

Goal state: the state to stop the search

Solutions: the set of paths that leads from the initial to the final goal state

Agent for route planner

Example of state space

Search tree

Terminology

Node

Parent

Child

Root: nodes with no parent

leaf: nodes with parent

Depth ($d$, $m$ is the max depth)

number of node between the root and the current node +1

Branching factor ($b$)

average number of children nodes that each node have

Node expansion

generate a node in the successor