Click or drag to resize
Accord.NET (logo)

QLearningUpdateState Method

Update Q-function's value for the previous state-action pair.

Namespace:  Accord.MachineLearning
Assembly:  Accord.MachineLearning (in Accord.MachineLearning.dll) Version: 3.6.0
Syntax
public void UpdateState(
	int previousState,
	int action,
	double reward,
	int nextState
)
Request Example View Source

Parameters

previousState
Type: SystemInt32
Previous state.
action
Type: SystemInt32
Action, which leads from previous to the next state.
reward
Type: SystemDouble
Reward value, received by taking specified action from previous state.
nextState
Type: SystemInt32
Next state.
See Also