An AI system consists of intelligent agents and their environments. An agent is anything which makes decisions as a person, firm or a software. Agents perceive their environment with the help of sensors. It is an autonomous entity which can act upon an environment using sensors and actuators to achieve goals.
Components of Intelligent Agents
The task of an intelligent agent is to design an agent program which implements the agent function. In other words, an agent comprises of an architecture and an agent program. By architecture, we mean the machinery on which an AI agent executes. Agent function maps a percept to an action. Agent program is an implementation of agent function. An agent program executes on the physical architecture to produce a function.
Examples of Agents
Intelligent Agents can interact with the environment in two ways: either by perception or action. Some examples of agents are:
Software agent – A piece of software that can work autonomously and interact with its environment
Robotic agent – It uses sensors to gain information about its surroundings.
Human agent – They have sensory organs to sense the environment
Types of Intelligent Agents
Intelligent Agents is classified into four classes based on their degree of perceived intelligence and capability :
Simple Reflex Agent:
This agent works on Condition-action rule, which suggests it maps the present state to action. It ignores percepts history and focuses only on the current percept during decision-making.
Model-based Reflex Agent:
This agent can add a partially observable environment and track things also. The agent keeps track of the internal state which includes adjustment by other percepts. This requires dependency on percept history.
Goal-based Agent:
In some cases, just knowing the information is not sufficient. The intent is to reduce distance from the goal, i.e., desired situations. Knowledge about the goal is an important factor. They are used to expand the capabilities based on the goal information. It may not be that efficient, but it is flexible.
Utility Agent:
These are similar to goal-based agent but also provides an extra utility measurement. A utility function helps in specifying trade-offs when the goals are conflicting. This comes handy when there are multiple possible alternatives. Besides, it can map each state to a real number for efficiency check of each action.
Learning Agent:
This type of agent learns from its past experience. It is called learning agent due to its learning capabilities. This can be further subdivided into:
- Learning element – This component learns from the environment and makes improvements accordingly
- Critic – learning elements can learn, analyze performance, and also take proper feedback from the critic to enhance the performance
- Performance element – This is responsible for selecting external action
- Problem generator – The problem generator suggests new actions which are informative.
Thus, we see that the objective of intelligent agents is to interact with the environment via perception and action. The software or devices having sensors helps in capturing information and processing data.
