CEpsilonGreedyDistribution Class Reference
Class for the
epsilon greedy action distribution. More...
#include <cpolicies.h>
Inheritance diagram for CEpsilonGreedyDistribution:
List of all
members.
Detailed Description
Class for the epsilon greedy action distribution.
This class implements the epsilon greedy action distribution.
Epsilon greedy policies take the greedy (best rated) action with
probability (1 - epsilon) and a random action with probability
epsilon. If there are more than one greedy action, always the first
action will be taken. To set epsilon please use the parameter
"EpsilonGreedy" or the constructor of the class. Its understood
that this distribution is not differentiable.
The class CEpsilonGreedyDistribution has following
Parameters:
- "EpsilonGreedy" : epsilon
Constructor & Destructor Documentation
|
CEpsilonGreedyDistribution::CEpsilonGreedyDistribution
|
( |
double |
epsilon |
) |
|
|
Member Function Documentation
| virtual void
CEpsilonGreedyDistribution::getDistribution
|
( |
CStateCollection *
|
state,
|
|
|
CActionSet *
|
availableActions,
|
|
|
double * |
values |
|
) |
[virtual] |
|
| |
Returns the distribution of the actions that is sampled by an
stochastic policy.
The function gets as input the current state, all available
actions, and the Q-Values (actually it can be any kind of value,
rating an action) of the actions as a double array. Usually only
this Q-Values are used for the distribution (the state is only used
for special exploration policies). The function has to overwrite
the Q-Values in double array with the distribution values.
Implements CActionDistribution.
|
The documentation for this class was generated from the following
file:
|
