Table 1 Summary of deep RL components.
From: Designing mechanically tough graphene oxide materials using deep reinforcement learning
DRL component | Notation | Description |
|---|---|---|
State | st | Current functional group locations on the GBP at time step t |
State space | S | All possible functional group locations on the GBP |
Action | at | To assign a functional group to a functional group spot on the GBP |
Action space | \({{{\mathcal{A}}}}({{{\mathbf{s}}}}_t)\) | All available functional group spots left given st |
Reward | rt | Standardized toughness if at terminal step; otherwise, 0 |
