Artificial Intelligence - UNIT - 1 Topic : 4. Agents and Environments

 

AENTS AND ENVIRONMENTS

---

Part A: Agents in Artificial Intelligence

---

1. Introduction

In Artificial Intelligence, an agent is anything that can perceive its environment using sensors and act upon that environment using actuators. Agents are designed to perform tasks autonomously, and they form the basic building blocks of intelligent systems.

---

2. What is an Agent?

An agent is defined as:

“An entity that perceives its environment through sensors and acts upon it using actuators to achieve specific goals.”

---

3. Components of an Agent

Component	Description
Sensors	Devices used to collect input from the environment
Actuators	Devices used to perform actions in the environment
Agent Program	The logic or code that decides what action to perform
Architecture	The platform (hardware/software) the agent runs on

---

4. Example of an Agent

Example: Self-Driving Car

• Sensors: Camera, GPS, radar
• Actuators: Steering, brake, throttle
• Agent Program: AI that decides how to drive safely
• Architecture: Embedded computer system in the car

---

Part B: Environments in AI

---

1. What is an Environment?

The environment is everything the agent interacts with. It provides input to the agent and receives the agent’s output actions.

The agent operates within an environment to achieve its goals.

---

2. Types of Environments

Environment Type	Description
Fully Observable	The agent has complete access to the environment's state (e.g., Chess)
Partially Observable	The agent has limited access to the state (e.g., Driving in fog)
Deterministic	The next state is completely predictable (e.g., Calculator)
Stochastic	Outcomes include randomness (e.g., Stock Market)
Static	The environment doesn’t change while the agent is thinking (e.g., Puzzle)
Dynamic	The environment changes over time (e.g., Autonomous Car)
Discrete	The environment has a finite number of actions (e.g., Board Games)
Continuous	Infinite states and actions are possible (e.g., Robot movement)
Single-Agent	Only one agent operates (e.g., Path-finding robot)
Multi-Agent	Multiple agents interact (e.g., Football game, Trading bots)

---

3. Agent-Environment Interaction

Agents and environments work in a loop:

• Agent receives percepts from the environment via sensors.
• Agent uses its program to decide an action.
• Agent performs the action through actuators.
• The environment changes based on this action.

Diagram (conceptual):

Agent → Sensors → [Percept] → Agent Program → [Action] → Actuators → Environment

---

4. PEAS Description (Task Environment)

To define the environment properly, we use PEAS:

PEAS Component	Description
Performance Measure	Criteria to judge success (e.g., speed, safety)
Environment	Everything the agent interacts with
Actuators	Devices that carry out actions
Sensors	Devices that perceive the environment

---

Example: Vacuum Cleaner Agent

PEAS Element	Description
Performance	Cleanliness, energy efficiency
Environment	Floors, dirt, walls
Actuators	Motor (move, suck dirt)
Sensors	Dirt sensor, bump sensor

---

Summary

• An agent is an autonomous entity that perceives and acts.
• The environment is the world where the agent operates.
• Environments can be fully/partially observable, static/dynamic, and deterministic/stochastic.
• Understanding both agents and environments is essential for designing smart AI systems.
• The PEAS framework helps define a complete task environment for AI design.