Agent-Based Systems Lecture 1

Study Notes

Lecture presented by Professor Dr. Magdy Zakarya and Dr. Dina Saif
Topics covered include agents and environments, rationality, PEAS (Performance measure, Environment, Actuators, Sensors), environmental types, and agent types
Agents are entities that perceive their environment through sensors and act upon it through actuators
Example human agents have eyes, ears, hands, and legs as sensors and actuators
Example robotic agents use cameras and infrared range finders as sensors and various motors as actuators
The agent function maps from percept histories to actions. The notation describes this as [f: P* → A]
The agent program runs on the physical architecture to produce the function.
Agent = architecture + program
Vacuum-cleaner world example with percepts (location and contents), actions (Left, Right, Stuck, NoOp), and the agent using a lookup table.

Rationality is measured by success
Agents' prior knowledge of the environment
Actions the agent can perform
The agent's percept sequence to date
A rational agent selects an action to maximize its performance measure based on percept evidence and built-in knowledge

Rationality differs from perfection, as it maximizes expected rather than actual outcomes
Percepts in situations might not be comprehensive, such as in games of cards or in situations where knowledge of other agents is unknown
A rational agent chooses the most likely optimal action given the available data and knowledge

An agent's autonomy is the extent to which its behavior is determined by experience, not the designer's knowledge.
Extremes of autonomy include having no autonomy (the environment/information is ignored) and complete autonomy (acting randomly/no program).
A desirable middle ground for autonomous agents exists.

PEAS (Performance measure, Environment, Actuators, Sensors): a framework to define intelligent agent designs
Specific examples given include:
- Automated Taxi Driver:
  - Performance: Safe, fast, legal, comfortable trip, maximizing profit
  - Environment: Roads, other traffic, pedestrians, and customers
  - Actuators: Steering wheel, accelerator, brake, signals, and horn
  - Sensors: Cameras, sonar, speedometer, GPS, odometer, and engine sensors
- Part-picking Robot:
  - Performance: Percentage of parts in correct bins
  - Environment: Conveyor belt with parts, and bins
  - Actuators: Jointed arm and hand
  - Sensors: Camera, and joint angle sensors
- Interactive English Tutor:
  - Performance: Maximizing student's test scores
  - Environment: Set of students
  - Actuators: Screen displays (exercises, suggestions, corrections)
  - Sensors: Keyboard and microphone