Sample within a uniform distribution 🔗

Samplings can be performed at random within a domain following a uniform ditribution, via the UniformDistribution(maximum) command. This task generates values uniformly distributed between zero and the maximum argument:
In the following example, 100 values are generated at random, uniformly distributed between 0 and 20.

val my_input = Val[Double]
val my_model = EmptyTask() set( (inputs, outputs) += my_input)

val exploration =
  DirectSampling(
    evaluation = my_model hook display,
    sampling= my_input in UniformDistribution[Double](max=20).take(100)
  )

exploration

Sample within a skewed uniform distribution 🔗

Custom domains can be defined using transformations on the uniform distribution. For instance in this next example, 100 values uniformly distributed between 0 and 20 are still generated at random, however, each one is then shifted by -10 through the map function. Thus, the sampling will be comprised of 100 values uniformly distributed between -10 and 10.

val my_input = Val[Double]
val my_model = EmptyTask() set( (inputs, outputs) += my_input)

val exploration =
  DirectSampling(
    evaluation = my_model hook display,
    sampling= my_input in UniformDistribution[Double](max=20).take(100).map(x => x -10)
  )

exploration

For more information on the @code{map} function, or other transformation functions, see the Language section.