A Novel Reinforcement Learning Architecture for Continuous State and Action Spaces

Figure 2

The A²C architecture.