PAGE Description

What is PAGE?

PAGE is an acronym used to define the properties of an agent's task environment:

• P: Percepts
• A: Actions
• G: Goals
• E: Environment

---

The PAGE Framework

1. Percepts

The inputs that the agent's sensors can detect at any given moment.

Example: For a taxi driver, percepts include video from cameras, GPS coordinates, and speedometer readings.

2. Actions

The possible set of operations the agent can perform.

Example: For a taxi driver, actions include steering, braking, accelerating, and signaling.

3. Goals

The objectives the agent is trying to achieve.

Example: For a taxi driver, goals include reaching the destination safely, minimizing time, and passenger comfort.

4. Environment

The external world in which the agent operates.

Example: For a taxi driver, the environment includes roads, traffic, pedestrians, and weather conditions.

---

PAGE Example: Automated Taxi

Component	Description
Percepts	Images, sonar, GPS, speedometer, engine sensors, keyboard
Actions	Steer, accelerate, brake, signal, speak to passenger
Goals	Safe, fast, legal, comfortable trip, maximize profits
Environment	Roads, other traffic, pedestrians, customers

---

Difference between PEAS and PAGE

While PEAS focuses on the interface (Sensors/Actuators) and performance, PAGE focuses more on the content and context of the agent's existence (Goals/Environment). Most modern AI textbooks prioritize PEAS.

---