Asynchronous Optimization via Random Search
Source:R/OptimizerAsyncRandomSearch.R
mlr_optimizers_async_random_search.RdOptimizerAsyncRandomSearch class that implements a simple Random Search.
Source
Bergstra J, Bengio Y (2012). “Random Search for Hyper-Parameter Optimization.” Journal of Machine Learning Research, 13(10), 281–305. https://jmlr.csail.mit.edu/papers/v13/bergstra12a.html.
Dictionary
This Optimizer can be instantiated via the dictionary
mlr_optimizers or with the associated sugar function opt():
Super classes
Optimizer -> OptimizerAsync -> OptimizerAsyncRandomSearch
Methods
Inherited methods
Examples
# example only runs if a Redis server is available
if (mlr3misc::require_namespaces(c("rush", "redux", "mirai"), quietly = TRUE) &&
redux::redis_available()) {
# define the objective function
fun = function(xs) {
list(y = - (xs[[1]] - 2)^2 - (xs[[2]] + 3)^2 + 10)
}
# set domain
domain = ps(
x1 = p_dbl(-10, 10),
x2 = p_dbl(-5, 5)
)
# set codomain
codomain = ps(
y = p_dbl(tags = "maximize")
)
# create objective
objective = ObjectiveRFun$new(
fun = fun,
domain = domain,
codomain = codomain,
properties = "deterministic"
)
# start workers
rush::rush_plan(worker_type = "mirai")
mirai::daemons(1)
# initialize instance
instance = oi_async(
objective = objective,
terminator = trm("evals", n_evals = 20)
)
# load optimizer
optimizer = opt("async_random_search")
# trigger optimization
optimizer$optimize(instance)
# all evaluated configurations
instance$archive
# best performing configuration
instance$archive$best()
# covert to data.table
as.data.table(instance$archive)
}
#> state x1 x2 y timestamp_xs
#> <char> <num> <num> <num> <POSc>
#> 1: finished -2.11085866 3.149893 -44.720348 2026-09-18 09:19:25
#> 2: finished -8.33732121 3.178398 -135.032815 2026-09-18 09:19:25
#> 3: finished -7.78396197 2.105379 -111.790803 2026-09-18 09:19:25
#> 4: finished 5.92632791 -4.231512 -6.932672 2026-09-18 09:19:25
#> 5: finished 0.98995613 2.678818 -23.269168 2026-09-18 09:19:25
#> 6: finished -3.50214144 1.111543 -37.178342 2026-09-18 09:19:25
#> 7: finished 2.77985531 4.933030 -53.541141 2026-09-18 09:19:25
#> 8: finished 9.86020590 0.052479 -61.100465 2026-09-18 09:19:25
#> 9: finished -4.77089430 4.984108 -99.590989 2026-09-18 09:19:25
#> 10: finished 8.62827623 3.264048 -73.172348 2026-09-18 09:19:25
#> 11: finished 5.20608164 -2.035465 -1.209287 2026-09-18 09:19:25
#> 12: finished -0.97588025 1.013553 -14.964468 2026-09-18 09:19:25
#> 13: finished -9.42569878 0.916582 -135.886207 2026-09-18 09:19:25
#> 14: finished 0.07085662 3.202987 -32.198639 2026-09-18 09:19:25
#> 15: finished -6.61173459 1.705343 -86.302224 2026-09-18 09:19:25
#> 16: finished 8.34573561 3.714461 -75.352353 2026-09-18 09:19:25
#> 17: finished -8.96971468 -3.263809 -110.404235 2026-09-18 09:19:25
#> 18: finished 1.89838435 2.698046 -22.478057 2026-09-18 09:19:25
#> 19: finished -2.71425870 1.813992 -35.398750 2026-09-18 09:19:25
#> 20: finished 1.30114551 -4.177370 8.125403 2026-09-18 09:19:25
#> state x1 x2 y timestamp_xs
#> <char> <num> <num> <num> <POSc>
#> worker_id timestamp_ys
#> <char> <POSc>
#> 1: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 2: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 3: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 4: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 5: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 6: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 7: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 8: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 9: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 10: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 11: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 12: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 13: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 14: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 15: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 16: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 17: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 18: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 19: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> 20: narrow_xinjiangovenator_33392330 2026-09-18 09:19:25
#> worker_id timestamp_ys
#> <char> <POSc>
#> keys x_domain_x1 x_domain_x2
#> <char> <num> <num>
#> 1: 9cd70165-d09a-4bf7-9e5e-a6996f4aaf22 -2.11085866 3.149893
#> 2: dfdc14ba-6e59-4c2f-ac12-5fbbc30bf88e -8.33732121 3.178398
#> 3: 846d4e7c-3e26-4833-94e4-3dd8194ecd95 -7.78396197 2.105379
#> 4: 847c401d-2f9b-46ee-9417-be67b35c1637 5.92632791 -4.231512
#> 5: 02c5159a-c6e0-4e9a-8eae-205a8ffafc0a 0.98995613 2.678818
#> 6: a376a821-4f63-4208-a7cd-a54852479c24 -3.50214144 1.111543
#> 7: 8e8b4621-0952-4f58-be9b-d57e6735a52b 2.77985531 4.933030
#> 8: 2cab44bb-3b8f-496e-a180-eac279078458 9.86020590 0.052479
#> 9: a4dc0d08-60e9-4cf6-b176-b532b5baea86 -4.77089430 4.984108
#> 10: 99c1106a-cb12-45f4-ba33-d4ee6a91d988 8.62827623 3.264048
#> 11: af661bd1-80a6-4de8-a8f0-3f5d1230004c 5.20608164 -2.035465
#> 12: 4677cd16-e27f-4df7-9baa-9535f0b82037 -0.97588025 1.013553
#> 13: 6ac8405b-1da6-4155-ac22-800b9cadcca3 -9.42569878 0.916582
#> 14: 8142252f-51e6-405d-999b-f93f8801c381 0.07085662 3.202987
#> 15: a981a0ce-a8be-4466-a9f3-d76470bfd6ec -6.61173459 1.705343
#> 16: fdfc7cf9-8e64-4ccf-88c5-65f67da6fda2 8.34573561 3.714461
#> 17: e90d364e-3d8f-4a9c-8168-8c939c92c815 -8.96971468 -3.263809
#> 18: e078fb93-e8f5-47fd-a056-8e7de89a7e0d 1.89838435 2.698046
#> 19: a4dd0227-db3b-43fd-8f44-c9fa11ec9a6f -2.71425870 1.813992
#> 20: cccbe98f-721a-4033-8944-922f7cc26ad9 1.30114551 -4.177370
#> keys x_domain_x1 x_domain_x2
#> <char> <num> <num>