On Jan 28, 9:31 pm, [EMAIL PROTECTED] wrote: > I also had a go at this problem for a bit of python practice, about 6 > months ago. I tried a few optimizations and my experience was that > with only 6 seeds, a hash table was very effective. So effective, in > fact, that it made all other optimizations more or less pointless.
My strategy was to walk through each solution "only once" (for a certain definition of "only once :), thus I was hoping not to need a hashtable. > Code below. Arguments are seeds first, then targets. Like this: > > C:\utils>countdown.py 7 8 50 8 1 3 923 > made target! > 50 * 8 = 400 > 400 - 1 = 399 > 399 / 3 = 133 > 133 * 7 = 931 > 931 - 8 = 923 > > Took 0.421 seconds Did you pick these numbers at random? Interestingly, the solution is unique: [0.27424502372741699] >>> list(test(nums=[7, 8, 50, 8, 1, 3], target=923, method=divide, >>> countdown=print_countdown)) (50*8-1)*7/3-8 (50*8-1)*7/3-8 [0.2748568058013916] To find just one solution: >>> list(test(nums=[7, 8, 50, 8, 1, 3], target=923, countdown=first_countdown)) >>> [0.11860203742980957] (list returned contains time taken to reach solution. By comparison your code takes 0.264s on my machine. Mine is faster on this example, but one example is not representative!) -- Arnaud -- http://mail.python.org/mailman/listinfo/python-list
