get_lower_boundary_design {adoptr} | R Documentation |
The optimization method minimize
is based on the package
nloptr
. This requires upper and lower boundaries for optimization.
Such boundaries can be computed via lower_boundary_design
respectively upper_boundary_design
.
They are implemented by default in minimize
.
Note that minimize
allows the user to define its own
boundary designs, too.
get_lower_boundary_design(initial_design, ...) get_upper_boundary_design(initial_design, ...) ## S4 method for signature 'OneStageDesign' get_lower_boundary_design(initial_design, n1 = 1, c1_buffer = 2, ...) ## S4 method for signature 'GroupSequentialDesign' get_lower_boundary_design(initial_design, n1 = 1, n2_pivots = 1, c1_buffer = 2, c2_buffer = 2, ...) ## S4 method for signature 'TwoStageDesign' get_lower_boundary_design(initial_design, n1 = 1, n2_pivots = 1, c1_buffer = 2, c2_buffer = 2, ...) ## S4 method for signature 'OneStageDesign' get_upper_boundary_design(initial_design, n1 = 5 * initial_design@n1, c1_buffer = 2, ...) ## S4 method for signature 'GroupSequentialDesign' get_upper_boundary_design(initial_design, n1 = 5 * initial_design@n1, n2_pivots = 5 * initial_design@n2_pivots, c1_buffer = 2, c2_buffer = 2, ...) ## S4 method for signature 'TwoStageDesign' get_upper_boundary_design(initial_design, n1 = 5 * initial_design@n1, n2_pivots = 5 * initial_design@n2_pivots, c1_buffer = 2, c2_buffer = 2, ...)
initial_design |
The initial design |
... |
optional arguments The values |
n1 |
bound for the first-stage sample size n1 |
c1_buffer |
shift of the early-stopping boundaries from the initial ones |
n2_pivots |
bound for the second-stage sample size n2 |
c2_buffer |
shift of the final decision boundary from the initial one |
initial_design <- TwoStageDesign( n1 = 25, c1f = 0, c1e = 2.5, n2 = 50, c2 = 1.96, order = 7L ) get_lower_boundary_design(initial_design)