From: rjh@cpax.org.uk   
      
   On 26/03/2022 10:28 pm, Malcolm McLean wrote:   
   > You have a list of candidates, ranked by score. You want to try them out in   
   order, with the better candidates being tried first. However you don't want   
   the process to be deterministic - each run should yield a separate order. And   
   you want even low-   
   ranked candidates to have some chance of being tried early.   
   >   
   > Is there a partial sort / partial shuffle which can achieve this?   
   >   
   > (The application is a crossword grid filler. I score the words, then try to   
   fit them into the grid. But I want a different grid each time, and I don't   
   want it to be too obvious that words with uncommon letters are never chosen.)   
      
   The obvious way is to change the score by awarding a pseudo-random   
   number of bonus points to the low rankers, maybe large to begin with but   
   reducing it as the fit proceeds so that it doesn't get tried too often.   
      
   --   
   Richard Heathfield   
   Email: rjh at cpax dot org dot uk   
   "Usenet is a strange place" - dmr 29 July 1999   
   Sig line 4 vacant - apply within   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)