Reinforcement Learning Sandbox
The AI Snake Lab codebase includes a large number of constants. This page outlines how these constants are named and organized. The files that contain the constants are in the ai_sim_lab/constants directory.
There are three types of constants, as shown in the table below:
| Constant Type | Example Name | Example Value |
|---|---|---|
| Fields | DModelRNN.MODEL |
"rnn_model" |
| Labels | DLabel.RNN_MODEL |
"RNN" |
| Defaults | DDef.MODEL |
DSim.RNN_MODEL |
Constants are further organized into categories as outlined in the table below:
| Category | Category File | Category Decription |
|---|---|---|
| Labels | DLabel.py |
Human readable labels for constants. |
| Defaults | DDef.py |
Default values for constants. |
| Fields | N/A | N/A |
As you can see from the table above, the files that contain the constant field definitions are not listed. See the next section, Constant Domains. below.
Constant fields are organized into domains.
| Domain File | Domain Description |
|---|---|
DColors.py |
Colors used in the TUI. |
DDir.py |
Directory names. These are a type of default constant. |
DEpsilon.py |
Constants related to the Epsilon Greedy algorithm. |
DFile.py |
File names. These are a type of default constant. |
DLayout.py |
Field definitions used in the TUI. |
DModelL.py |
Constants related to the Linear Model. |
DModelRNN.py |
Constants related to the RNN Model. |
DPlot.py |
Constants related to plotting. |
DReplayMemory.py |
Constants related to the Replay Memory. |
DSim.py |
Simulation wide constants. |