Date: 12/01/05 (Algorithms)    Keywords: no keywords

One of the most frustrations in my current project is it takes a long time in order to see if the modification creates any difference. Since heuristic randomness is involved, I need to take the average answer (e.g., trial -> running the program for 100 times / get the best & average). Each trial can take a longggggggggg period of time.

It's also frustrated that it's difficult to know what values, combination, settings etc would be good to use or to avoid.


What do you do in such case ?

Source: http://www.livejournal.com/community/algorithms/67896.html