| Interface | Description |
|---|---|
| Explorer<T> |
Exposes a method to choose an action given a generic context, and obtain the relevant
exploration bits.
|
| Class | Description |
|---|---|
| BootstrapExplorer<T> |
The Bootstrap explorer randomizes over the actions chosen by a set of default policies.
|
| EpsilonGreedyExplorer<T> |
The epsilon greedy exploration algorithm.
|
| GenericExplorer<T> |
GenericExplorer provides complete flexibility.
|
| MwtExplorer<T> |
The top-level MwtExplorer class.
|
| SoftmaxExplorer<T> |
In some cases, different actions have a different scores, and you would prefer to
choose actions with large scores.
|
| TauFirstExplorer<T> |
The tau-first explorer collects exactly tau uniform random exploration events, and then
uses the default policy thereafter.
|
| VariableActionBootstrapExplorer<T extends VariableActionContext> |
The Bootstrap explorer randomizes over the actions chosen by a set of default policies.
|
| VariableActionEpsilonGreedyExplorer<T extends VariableActionContext> |
The epsilon greedy exploration algorithm.
|
| VariableActionGenericExplorer<T extends VariableActionContext> |
GenericExplorer provides complete flexibility.
|
| VariableActionSoftmaxExplorer<T extends VariableActionContext> |
In some cases, different actions have a different scores, and you would prefer to
choose actions with large scores.
|
| VariableActionTauFirstExplorer<T extends VariableActionContext> |
The tau-first explorer collects exactly tau uniform random exploration events, and then
uses the default policy thereafter.
|
Copyright © 2016. All Rights Reserved.