mk_exponentially_weighted_acceptance_func {designit}

Alternative acceptance function for multi-dimensional scores with exponentially downweighted score improvements from left to right

Description

Alternative acceptance function for multi-dimensional scores with exponentially downweighted score improvements from left to right

Usage

mk_exponentially_weighted_acceptance_func(
  kappa = 0.5,
  simulated_annealing = FALSE,
  temp_function = mk_simanneal_temp_func(T0 = 500, alpha = 0.8)
)

Arguments

kappa

Coefficient that determines how quickly the weights for the individual score improvements drop when going from left to right (i.e. first to last score). Weight for the first score's delta is 1, then the original delta multiplied with kappa^(p-1) for the p'th score

simulated_annealing

Boolean; if TRUE, simulated annealing (SA) will be used to minimize the weighted improved score

temp_function

In case SA is used, a temperature function that returns the annealing temperature for a certain iteration number

Value

Acceptance function which returns TRUE if current score should be taken as the new optimal score, FALSE otherwise