Profiling Optimizations (like Galapagos)

Wish: having time stamps or duration recorded in the logdata from an optimization run (see list below, from Galapagos).

Why? : I find myself doing multiple optimization tests in order to find optimal settings, and most importantly, finding threshold values which would abort the process as early as possible, whether it be a finess value or number of stagnant generations. Time (duration) is often the most important factor but unfortunately I don’t find any timestamps in the log data.

// Rolf


Example: Galapagos log data without timestamps :anguished: :

Generation 2
{
  Bio-Diversity: 0.746
  Genome[0], Fitness=603.00, Genes [42% · 38% · 48% · 67%]
  Genome[1], Fitness=520.00, Genes [58% · 53% · 48% · 45%]
  {
    Record: Point Mutation at index 0: 0.9167 -> 0.89
  }
  Genome[2], Fitness=483.00, Genes [58% · 53% · 48% · 43%]
  {
    Record: Point Mutation at index 1: 0.0667 -> 0.0629
  }
  Genome[3], Fitness=201.00, Genes [40% · 52% · 53% · 53%]
  Genome[4], Fitness=33.00, Genes [22% · 58% · 72% · 55%]
  {
    Record: Point Mutation at index 0: 0.2833 -> 0.2692
    Record: Point Mutation at index 2: 0.2 -> 0.4273
  }
  Genome[5], Fitness=6.00, Genes [37% · 38% · 55% · 75%]
  Genome[6], Fitness=0.00, Genes [67% · 25% · 40% · 35%]
  ...
  ...  Aso.
}