Conclusions or Reasons why you might care

• Theory works. Fitness is a needle, mean=$\frac{1}{2}$ $\sigma=\frac{2^{-n/2}}{2\sqrt{m}}$
  Chance of reversible solution $2^{-\vert{\rm testset}\vert}=2^{-{m}2^{n}}$

  
  

• Possible applications of reversible computing (e.g. safety critical, ultra low power consumption) but no human experience with reversible programming. Automatic programming is possible and may be economic choice.

  
  

• Evolutionary search shown to out perform hill climbing.
  Random search provably useless.

  
  

• CCNOT competitive with usual GP function set. Could try reversible computing as an alternative on every day problems.

  
  

• Minimal solutions hurt

  There is one smallest solution.
  But exponentially many non-minimal solutions.