Custom Reinforcement Learning Environment

Custom Reinforcement Learning Environment

Arguments

step	[function(self, action)] Custom step function.
reset	[function(self)] Custom reset function.
visualize	[function(self)] Optional custom visualization function.
discount	[numeric(1) in (0, 1)] Discount factor.
action.names	[named integer] Optional action names for a discrete action space.

Usage

makeEnvironment("custom", step, reset, visualize = NULL, discount = 1, action.names = NULL)

Methods

• $step(action) Take action in environment. Returns a list with state, reward, done.

• $reset() Resets the done flag of the environment and returns an initial state. Useful when starting a new episode.

• $visualize() Visualizes the environment (if there is a visualization function).

Examples

step = function(self, action) {
  state = list(mean = action + rnorm(1), sd = runif(1))
  reward = rnorm(1, state[[1]], state[[2]])
  done = FALSE
  list(state, reward, done)
}

reset = function(self) {
  state = list(mean = 0, sd = 1)
  state
}

env = makeEnvironment(step = step, reset = reset)
env$reset()
#> $mean
#> [1] 0
#> 
#> $sd
#> [1] 1
#> 
env$step(100)
#> $state
#> $state$mean
#> [1] 98.59996
#> 
#> $state$sd
#> [1] 0.6007609
#> 
#> 
#> $reward
#> [1] 97.99559
#> 
#> $done
#> [1] FALSE
#>