Navigation
BlueprintAPI > BlueprintAPI/LearningAgents
Specifies a new exclusive discrete action. This represents an action which is an exclusive choice from a number of discrete options, sampled from a Categorical distribution.
Target is Learning Agents Actions
Inputs
| Type | Name | Description |
|---|---|---|
| object | Schema | The Action Schema |
| integer | Size | The number of discrete options in the action. |
| real | Prior Probabilities | The prior probabilities of each option. Can be left empty to use a uniform distribution over options. Should sum to one. |
| name | Tag | The tag of this new action. Used during action object validation and debugging. |
Outputs
| Type | Name | Description |
|---|---|---|
| struct | Return Value | The newly created action schema element. |