Run blockwise online design optimization decision support — run_gazepoint_online_design

Provides a safe, dependency-light decision-support/simulation helper for online design optimization. The function recommends the next condition by combining expected model-discrimination utility with optional exploration and balancing penalties.

Usage

run_gazepoint_online_design_optimization(
  candidate_table,
  condition_col = "condition",
  utility_col = "expected_utility",
  block_col = NULL,
  cost_col = NULL,
  previous_assignments = NULL,
  exploration_weight = 0.1,
  balance_weight = 0.1,
  maximise = TRUE
)

Arguments

candidate_table: A data frame containing candidate conditions.
condition_col: Candidate condition column.
utility_col: Expected utility/model-discrimination column.
block_col: Optional block column.
cost_col: Optional cost or burden column subtracted from utility.
previous_assignments: Optional previous condition assignments.
exploration_weight: Weight for favouring under-sampled conditions.
balance_weight: Weight for penalising over-sampled conditions.
maximise: Logical. If TRUE, select highest score.

Value

A list with overview, ranked_candidates, recommendation, assignment_summary, and settings.

Details

This function does not control stimulus presentation software and should not be used as autonomous real-time experiment control without separate ethical, preregistration, and software-integration review.