1. Introduction:

“ An agent is a system that decides for itself what it must do to achieve its objectives (determined at its design). ”

 2. Agent: definition & Characteristics:

An agent is a physical or virtual entity that has all or part of the following functionalities:

•       Is able to act in an environment,

•       Can communicate directly with other agents

•       Is motivated by a set of trends,

•       Has its own resources,

•       Is able to perceive his environment,

•       Has a partial representation of its environment

•       Has skills and offers services,

•       His behavior tends to satisfy his objectives according to his perception, his representations and the communications he receives.

  2.1 Properties of an agent:

• Autonomy : Behaviour is dependent on perceptions and representations of the environment. It acts without external control.
• Proactive: Takes initiatives to achieve its objectives, without waiting for another entity to intervene.
• Flexible: It adapts its behaviour to the environment and can collaborate with other agents to better achieve its objectives.
• Social: Interacts with other agents to achieve their goals or assist them in their activities.
• Located: It perceives and acts in its environment taking into account the space-time constraints.

Example :

A simple system, such as a thermostat, can be considered an intelligent agent.

Why is the thermostat a smart agent?

•       Perception: It measures the ambient temperature using a sensor.

•       Decision: Based on the measured temperature, it compares this value to a user-defined target temperature.

•       Action: If the temperature is too low, it turns on the heating;  if it is too high, it turns off.

Examples of agents:

• Printer (Mechanical Agent):

• Biological Agents (Plants, Animals, Humans):

• Software Agents (Programs)

2.2 Agent Types:

- Autonomous Agents :An autonomous agent is a system capable of making decisions independently based on its environment and its objectives (e.g. Smart cars).

- Reactive Agents

A reactive agent is a type of agent that reacts directly to changes in its environment without complex planning or internal modeling (e.g. printer).

- Hybrid Agents

A hybrid agent combines the characteristics of reactive agents and autonomous (deliberative) agents (e.g. robot cleaner).

2.3 Properties of an environment:

• Accessible vs. Inaccessible:

       - An accessible environment: provides complete and up-to-date  information on its    status. (e.g.Chess game)

       - An environment inaccessible: does not allow to obtain all the  information                 (e.g.Autonomous driving in the city).

• Deterministic vs. Non-Deterministic:  

  -  Deterministic  : Each action produces a certain effect, without  uncertainty. (e.g. Sudoku games)

  -  Non-deterministic  : An action may fail or produce unpredictable results. (e.g. Robot that delivers outside)

• Discreet vs Continuous:  

  -  Discrete  : Finite number of possible actions (e.g., chess)

  - Continuous  states change smoothly (e.g. taxi speed and  position). 

• Single vs Multi-Agent:  

  -  Simple  : Only one agent acts (e.g. solve a puzzle alone). 

  -  Multi-Agent  Multiple agents interact (e.g. play chess).

•  Static vs. Dynamic:  

  -  Static  : The environment does not change without  agent intervention (e.g., puzzle). 

  -  Dynamic  : The environment can change independently of the agent (e.g., taxi driving).