sampling

type AverageStrategyParams ¶

type AverageStrategyParams struct {
	Epsilon float32 // 0 < e <= 1
	Tau     float32 // t >= 1
	Beta    float32 // b >= 0
}

type AverageStrategySampler ¶

type AverageStrategySampler struct {
	// contains filtered or unexported fields
}

AverageStrategySampler implements cfr.Sampler by sampling some player actions according to the current average strategy strategy.

func NewAverageStrategySampler(params AverageStrategyParams) *AverageStrategySampler

func (as *AverageStrategySampler) Sample(node cfr.GameTreeNode, pol cfr.NodePolicy) []float32

type ExternalSampler ¶

type ExternalSampler struct {
	// contains filtered or unexported fields
}

ExternalSampler implements cfr.Sampler by sampling all player actions.

func NewExternalSampler() *ExternalSampler

func (es *ExternalSampler) Sample(node cfr.GameTreeNode, policy cfr.NodePolicy) []float32

type MultiOutcomeSampler ¶

type MultiOutcomeSampler struct {
	// contains filtered or unexported fields
}

MultiOutcomeSampler implements cfr.Sampler by sampling at most k player actions with probability according to the current strategy.

func NewMultiOutcomeSampler(k int, explorationEps float32) *MultiOutcomeSampler

func (os *MultiOutcomeSampler) Sample(node cfr.GameTreeNode, policy cfr.NodePolicy) []float32

type OutcomeSampler ¶

type OutcomeSampler struct {
	// contains filtered or unexported fields
}

OutcomeSampler implements cfr.Sampler by sampling one player action according to the current strategy.

func NewOutcomeSampler(explorationEps float32) *OutcomeSampler

func (os *OutcomeSampler) Sample(node cfr.GameTreeNode, policy cfr.NodePolicy) []float32

type RobustSampler ¶

type RobustSampler struct {
	// contains filtered or unexported fields
}

RobustSampler implements cfr.Sampler by sampling a fixed number of actions uniformly randomly.

func NewRobustSampler(k int) *RobustSampler

func (rs *RobustSampler) Sample(node cfr.GameTreeNode, policy cfr.NodePolicy) []float32