02.1 Agent Architecture

Architecture

blueprints of the agent’s brain

Architecture

defines how it processes info, decides, and then acts

Simple Reflex Agents

act only on current percepts

Simple Reflex Agents

no memory, no context

Model Based Agents

use internal state or memory of the world

Model Based Agents

handle partially observable environments

Goal Based Agents

plan actions to reach goal

Utility Based Agents

evaluates outcomes and picks the best move

Utility Based Agents

balances trade-offs

Learning Agents

adapt from experience

Learning Agents

improve performance over time

Environment

the world where agents operate

Environment

provides percepts + reacts to actions

Environment

can be hostile, friendly, or neutral

Fully vs. Partially Observable

environment properties (1)

Deterministic vs. Stochastic

environment properties (2)

Episodic vs. Sequential

environment properties (3)

Static vs. Dynamic

environment properties (4)

Discrete vs. Continuous

environment properties (5)

Single Agent vs. Multi Agent

environment properties (6)