Parallelised Q-Network for Continuous Action Spaces

Extension of PQN to continuous action spaces.

Simplifying Deep Temporal Difference Learning

A modern implementation of Deep Q-Network without target networks and replay buffers.