In Artificial Intelligence agents, There are four types of intelligence agents. The classification is based on the level of acquired potential.
The first type of Artificial Intelligence agent is the:
Simple Reflex Agent
It is the most basic type of agent that executes its action from the present state by or through the pre-determined rules for the given conditions.
A simple reflex agent quickly scans its knowledge base if something happens in its environment. It responds to the situation after it scanned the knowledge base which contains the predetermined rules.
This type of Intelligence agent is so basic and simple that it cannot compute complex equations or solve complicated problems and are fully-observable in the current happenings in its environment. In short, it is built on the condition-action rule.
Example: electronic door opening when the right keycard is swiped this explains the mechanism of a simple reflex agent very well.
Model-Based Reflex Agent
Model-based agents use Internal memory to store their navigation history and then use that to analyze things about their current environment though everything they need to know cannot be immediately observed. They use their internal memory and precept history to make decisions about an internal model. Roomba vacuum cleaner is an example of this.
A goal-based agent is a type of Intelligence agent that is flexible, they have the knowledge used for decision making and they can present in a definite way and it can be modified. A goal-based agent does a lot more work than the reflex agent.
Goal-based agents have a set of goals and they can use these goals with a set of actions and their outcomes to assess which action achieves the goal which may take one action or more to achieve.
Google’s Waymo is an example of a Goal-based agent they are programmed with a goal to deliver the passenger where they intended to go.
From the word utility which means make use of, the Utility-based agent is the type of Intelligence agent that makes its usefulness and it is its usefulness which makes it distinct from the other Intelligence agent. Utility-based agents act based on what the goal is like the goal-based agent but:
For example, a goal-based agent only takes you to the place you want to be in by setting the goal to it while a utility-based agent will find another route for you if you encounter problems along the road.
A learning agent is a type of Intelligence agent that can perform tasks on its own like analyzing performance and even look for new ways to improve the tasks. It is learning from its experiences.
For example, at school, you will be asked to take the test and the test would be marked as it will be critiqued. The teacher will assess the test, mark, and see what could be improved and then guide you on how to do better next time. In this example, the Teacher is the Learning element and you as the Student is the performance element.