Reinforcement Learning

Reinforcement learning is a class of machine learning in which an agent learns to make sequences of decisions by trial and error, receiving rewards or penalties for its actions. It is applied to control and optimisation problems such as energy management and process tuning.

Rather than learning from labelled examples, a reinforcement-learning agent interacts with an environment, takes actions, observes the resulting state and reward, and gradually learns a policy that maximises cumulative reward. In industry it shows promise for complex control and scheduling where good actions are hard to specify directly. Because real-world trial and error can be costly or unsafe, training is usually done against a simulation or digital twin before deployment.

Related terms

Advanced Process Control (APC) · Digital Twin · Machine Learning (Industrial)