Skip to contents

Epsilon-greedy action distribution, deterministic LCG draw (hmeg)

Usage

morie_geron_epsilon_greedy_alt(Q, s, epsilon, seed = 0)

Arguments

Q

Table (S, A) or (A,).

s

State index (0-based).

epsilon

Rate in [0, 1].

seed

LCG seed.

Value

List with action, probabilities, greedy_action, greedy_actions, is_exploratory.

Examples

morie_geron_epsilon_greedy_alt(matrix(c(1, 5, 2), 1, 3), s = 0,
    epsilon = 0.3)
#> $action
#> [1] 1
#> 
#> $probabilities
#> [1] 0.1 0.8 0.1
#> 
#> $greedy_action
#> [1] 1
#> 
#> $greedy_actions
#> [1] 1
#> 
#> $q_values
#> [1] 1 5 2
#> 
#> $is_exploratory
#> [1] FALSE
#> 
#> $epsilon
#> [1] 0.3
#> 
#> $estimate
#> [1] 0.8
#> 
#> $n
#> [1] 3
#> 
#> $method
#> [1] "epsilon-greedy action distribution with deterministic LCG sampling"
#>