public void EpsilonGreedyFixedActionUsingVariableActionInterface()
{
int numActions = 10;
float epsilon = 0f;
var policy = new TestPolicy <VariableActionTestContext>();
var testContext = new VariableActionTestContext(numActions);
var explorer = new EpsilonGreedyExplorer(epsilon);
EpsilonGreedyWithContext(numActions, testContext, policy, explorer);
}
public void GenericFixedActionUsingVariableActionInterface()
{
int numActions = 10;
var scorer = new TestScorer <VariableActionTestContext>(1, numActions);
var testContext = new VariableActionTestContext(numActions)
{
Id = 100
};
var explorer = new GenericExplorer();
GenericWithContext(numActions, testContext, explorer, scorer);
}
public void TauFirstFixedActionUsingVariableActionInterface()
{
int numActions = 10;
int tau = 0;
var testContext = new VariableActionTestContext(numActions)
{
Id = 100
};
var policy = new TestPolicy <VariableActionTestContext>();
var explorer = new TauFirstExplorer(tau);
TauFirstWithContext(numActions, testContext, policy, explorer);
}
public void SoftmaxFixedActionUsingVariableActionInterface()
{
int numActions = 10;
float lambda = 0.5f;
int numActionsCover = 100;
float C = 5;
var scorer = new TestScorer <VariableActionTestContext>(1, numActions);
var explorer = new SoftmaxExplorer(lambda);
int numDecisions = (int)(numActions * Math.Log(numActions * 1.0) + Math.Log(numActionsCover * 1.0 / numActions) * C * numActions);
var contexts = new VariableActionTestContext[numDecisions];
for (int i = 0; i < numDecisions; i++)
{
contexts[i] = new VariableActionTestContext(numActions)
{
Id = i
};
}
SoftmaxWithContext(numActions, explorer, scorer, contexts);
}