QLearner Class

Description:

  A class for creating QLearner objects.

Usage:

local ML = require("ML")
local qLearner = ML.QLearner()

pack

Type: Function.

Description:

  Construct a state from given hints and condition values.

Signature:

pack: function(self: QLearnerClass, hints: {integer}, values: {integer}): --[[state]] integer

Parameters:

Parameter	Type	Description
hints	{integer}	Representing the max number of possible hints. For example, if there are two conditions, and each condition has 3 possible values (0, 1, 2), then the hints array is {3, 3}.
values	{integer}	The condition values as discrete values.

Returns:

Return Type	Description
integer	The packed state value.

unpack

Type: Function.

Description:

  Deconstruct a state from given hints to get condition values.

Signature:

unpack: function(self: QLearnerClass, hints: {integer}, state: integer): {integer}

Parameters:

Parameter	Type	Description
hints	{integer}	Representing the max number of possible hints. For example, if there are two conditions, and each condition has 3 possible values (0, 1, 2), then the hints array is {3, 3}.
state	integer	The state integer to unpack.

Returns:

Return Type	Description
{integer}	The condition values as discrete values.

__call

Type: Metamethod.

Description:

  Create a new QLearner object with optional parameters for gamma, alpha, and maxQ.

Signature:

metamethod __call: function(
		self: QLearnerClass,
		gamma?: number --[[0.5]],
		alpha?: number --[[0.5]],
		maxQ?: number --[[100.0]]
	): QLearner

Parameters:

Parameter	Type	Description
gamma	number	The discount factor for future rewards. Defaults to 0.5.
alpha	number	The learning rate for updating Q-values. Defaults to 0.5.
maxQ	number	The maximum Q-value. Defaults to 100.0.

Returns:

Return Type	Description
QLearner	The newly created QLearner object.