prevision_quantum_nn.applications.reinforcement_learning.q_learning¶
Q-Learning module
Module Contents¶
Classes¶
Q-table Learner |
-
class
prevision_quantum_nn.applications.reinforcement_learning.q_learning.QLearner(params)¶ Bases:
prevision_quantum_nn.applications.reinforcement_learning.base_learner.BaseLearnerQ-table Learner
This class implements a Q-table.
-
params¶ containing the parameters of the model
- Type
dictionary
-
get_cell(self, state)¶ Returns the cell in which the state is.
- Parameters
state (numpy array) – state at which we wish to obtain the corresponding cell
- Returns
- tuple
cell indices at which the Q-table will be updated
- Return type
cell
-
fit(self, state, reward)¶ Fits the Q-table at a given state.
- Parameters
state (numpy array) – state at which the Q-table needs to fit
reward (float) – reward that was yielded when passing through this state
-
forward(self, state, action=None)¶ Forwards the Q-able through the state (action) provided.
- Parameters
state (numpy array) – state at which the Q-table needs to be evaluated
action (int) – if the Q-table is formed with the state action method, action needs to be provided as an option.
- Returns
- float
the actual Q-value at the state provided
- Return type
value
-