What Is an Intelligent Agent?

An intelligent agent is a computer entity that functions independently and works to complete pre-written goals. To complete goals, the intelligent agent must be able to analyze and adapt to outside stimuli; otherwise, it may be unable to accurately achieve goals. While some simple systems have one agent, complex systems come with sub-agents that function below the main agent to perform a list of minor tasks and report directly to the main agent. The agent is commonly able to learn through artificial intelligence, but the agent is made only to have certain responses to outside stimuli, and it cannot form new responses.

In computer science, an intelligent agent is a section of a system that is made to function independently, and it commonly is made with artificial intelligence. Along with this, the agent is programmed to have certain goals, depending on what the system is supposed to do. For example, if the system is an air conditioner, then goals can include cooling down the air and turning on and off when needed. While the system is able to learn to be more efficient, it cannot go against the pre-written goals.

To achieve these goals, the intelligent agent must be capable of analyzing itself and adapting to situations. With an air-conditioning system, the device is made to cool the air, so it must be able to analyze the outside temperature. By analyzing the outside temperature and its own system, the agent will know when it is appropriate to turn on or off or to adapt to the situation. This also enables the agent to check for errors in its logic, which it then can correct to better serve the pre-written goals.

Simple systems usually have one intelligent agent that can easily control all the functions. Complex systems may require several intelligent agents, but they are generally given an hierarchy to keep from internal logic struggles. Sub-agents perform smaller tasks and are typically governed by a main agent that supervises these tasks and ensures the system runs correctly.

During its operation, an intelligent agent will typically learn how to best serve its purpose by checking error and success rates. While the agent can learn to perform actions better, it can only perform pre-written actions. An air conditioner cannot perform a task that is not written into its code, such as moving itself to better cool a room. Much like the goals, it cannot change its actions, but it can change how well the system performs these actions.