# DQN

A reinforcement-learning algorithm that approximates the action-value function with a deep network, using replay memory and a target network.

## What it is used for

Discrete-action control problems such as dispatching and scheduling decisions.