Actor-Critic Algorithm (Glossary)

The actor-critic algorithm is a term used in the field of artificial intelligence and automation. It is one of the advanced methods used in reinforcement learning, in which computers or robots learn to make decisions independently.

The Actor-Critic Algorithm has two important components: the "Actor" and the "Critic". The actor decides which action should be carried out next. The Critic then evaluates how good or bad this decision was. They work together to continuously improve the behaviour of the system through learning.

A practical example: Imagine a warehouse robot that is supposed to sort goods efficiently. The Actor selects which shelf the robot goes to next. The Critic then checks whether this choice has led to faster work processes. If the result is positive, the algorithm reinforces such decisions for the future.

Thanks to the Actor-Critic Algorithm, machines can develop better strategies in complex and changing environments - for example in modern, automated factories or autonomous vehicles. This improves efficiency and flexibility and makes systems self-learning and customisable.