Diversity

1

• Diversity
• State of the art
• Chemical diversity ? Objective space diversity
  (first sometimes/often follows from last)
• Best measure context dependent
• Too many solutions may be bad, as are too few
• Diversity hedge against uncertainty (don't
  know/ can't find single best solution)
• Crowdingdistance
• Hypervolumes
• Chemical diversity beneficial for
  patents/side-effect profiles
• What can be better
• Better sampling of the Pareto front in case of
  large search space better GA or other algorithms
• Promising techniques to achieve it
• Rudolph Preuss find different set from
  search space that leads to same solution in
  objective space
• niching related emoa