Artificial Intelligence - UNIT - 1 Topic - 7 : Structure of Agents

 Part A: Understanding Agent Structure in Artificial Intelligence

---

1. Introduction

In Artificial Intelligence, an agent is an entity that perceives its environment and takes actions to achieve its goals. But how does it decide what to do?

The structure of an agent defines how it is internally designed to make decisions, process inputs, and select actions. It is like the brain and body of the agent.

---

2. What is Agent Structure?

Agent structure refers to the internal architecture or design of an agent — including how it processes percepts (inputs), stores knowledge, makes decisions, and performs actions.

It determines how the agent reacts to the environment.

---

3. Components of an Agent Structure

Component	Description
Sensors	Collect information (percepts) from the environment
Actuators	Perform actions in the environment
Agent Program	Software or logic that decides what action to take
Architecture	The platform (hardware/software) on which the agent operates

---

4. Types of Agent Structures

There are different types of agent designs based on how simple or intelligent the agent is. Each structure suits different kinds of environments and tasks.

---

 1. Simple Reflex Agents

• React only to current input using predefined rules.
• Do not use history or memory.
• Use “if condition, then action” logic.

Example:
A thermostat turns on heating if the temperature is below 20°C.

IF temperature < 20°C THEN turn on heater

---

2. Model-Based Reflex Agents

• Have a model (memory) of the world to track what’s happening.
• Can handle partially observable environments.
• Use internal state to remember the past.

Example:
A smart vacuum remembers which rooms it has cleaned.

---

 3. Goal-Based Agents

• Decide actions by comparing future outcomes based on a specific goal.
• Involve planning and search algorithms.
• Better decision-making than reflex agents.

Example:
A robot in a maze tries to reach the exit using path planning.

---

4. Utility-Based Agents

• Choose actions that maximize their utility (happiness/satisfaction).
• Utility = How beneficial or useful the outcome is.
• Can handle multiple goals and choose the best.

Example:
A delivery robot chooses the shortest and safest route to save time and battery.

---

5. Learning Agents

• Can learn from experience and improve their performance over time.
• Adapt to new environments.
• Have components like:
• Learning element – Improves agent
• Critic – Gives feedback
• Performance element – Chooses actions
• Problem generator – Suggests new experiences

Example:
A self-driving car gets better at driving after each trip by learning from traffic data.

---

5. Comparison of Agent Structures

Agent Type	Uses Memory	Has Goals	Learns	Example
Simple Reflex Agent	❌	❌	❌	Light turns on when someone enters
Model-Based Agent	✅	❌	❌	Smart vacuum
Goal-Based Agent	✅	✅	❌	Maze-solving robot
Utility-Based Agent	✅	✅	❌	Route-optimizing delivery robot
Learning Agent	✅	✅	✅	Self-driving car

---

6. Summary

• The structure of agents defines how an AI system works internally to make decisions.
• Different agents are built based on task complexity, environment, and goals.
• Advanced agents not only think logically but also learn and improve.
• Understanding agent structure helps in building intelligent and adaptive AI systems.