Get Fitness Values from OptimInstance

Get fitness values in the correct form as used by Selector operators from an OptimInstance. This works for both single-criterion and multi-criterion optimization, and entails multiplying objectives with -1 if they are being minimized, since Selector tries to maximize fitness.

mies_get_fitnesses(inst, rows)

Arguments

inst

(OptimInstance)
Optimization instance to evaluate.

rows

optional (integer)
Indices of rows within inst to consider. If this is not given, then the entire archive is used.

Value

numeric

matrix with length(rows) (if rows is given, otherwise nrow(inst$archive$data)) rows and one column for each objective: fitnesses to be maximized.

See also

Other mies building blocks: mies_evaluate_offspring(), mies_generate_offspring(), mies_init_population(), mies_select_from_archive(), mies_step_fidelity(), mies_survival_comma(), mies_survival_plus()

Examples

set.seed(1)
library("bbotk")
lgr::threshold("warn")

# Define the objective to optimize
objective <- ObjectiveRFun$new(
  fun = function(xs) {
    z <- exp(-xs$x^2 - xs$y^2) + 2 * exp(-(2 - xs$x)^2 - (2 - xs$y)^2)
    list(Obj = z)
  },
  domain = ps(x = p_dbl(-2, 4), y = p_dbl(-2, 4)),
  codomain = ps(Obj = p_dbl(tags = "maximize"))
)

# Get a new OptimInstance
oi <- OptimInstanceSingleCrit$new(objective,
  terminator = trm("evals", n_evals = 100)
)

mies_init_population(inst = oi, mu = 3)

oi$archive
#> <Archive>
#>        x     y   dob   eol  x_id     Obj           timestamp batch_nr
#>    <num> <num> <num> <num> <num>   <num>              <POSc>    <int>
#> 1: -0.41  3.45     1    NA     1 0.00075 2024-05-13 04:41:07        1
#> 2:  0.23 -0.79     1    NA     2 0.50760 2024-05-13 04:41:07        1
#> 3:  1.44  3.39     1    NA     3 0.21083 2024-05-13 04:41:07        1

mies_get_fitnesses(oi, c(2, 3))
#>            Obj
#> [1,] 0.5076046
#> [2,] 0.2108264

###
# Multi-objective, and automatic maximization:
objective2 <- ObjectiveRFun$new(
  fun = function(xs) list(Obj1 = xs$x^2, Obj2 = -xs$y^2),
  domain = ps(x = p_dbl(-2, 4), y = p_dbl(-2, 4)),
  codomain = ps(
    Obj1 = p_dbl(tags = "minimize"),
    Obj2 = p_dbl(tags = "maximize")
  )
)
# Using MultiCrit!
oi <- OptimInstanceMultiCrit$new(objective2,
  terminator = trm("evals", n_evals = 100)
)

mies_init_population(inst = oi, mu = 3)

oi$archive
#> <Archive>
#>        x     y   dob   eol  x_id  Obj1  Obj2           timestamp batch_nr
#>    <num> <num> <num> <num> <num> <num> <num>              <POSc>    <int>
#> 1:   3.7 -1.63     1    NA     1  13.5 -2.65 2024-05-13 04:41:07        1
#> 2:   2.0 -0.76     1    NA     2   3.9 -0.58 2024-05-13 04:41:07        1
#> 3:   1.8 -0.94     1    NA     3   3.1 -0.88 2024-05-13 04:41:07        1

# Note Obj1 has a different sign than in the archive.
mies_get_fitnesses(oi, c(2, 3))
#>           Obj1       Obj2
#> [1,] -3.860387 -0.5839291
#> [2,] -3.149504 -0.8848403