guillaume

c828545f3b 11.76M on input D (+760k) with a simple coeff change Of course i tried different coeffs before hitting a good one
96437b9064 11M on input D by changing scoring function
5c0828ac27 Revert "Some inputs perform better picking one of the next best rides" This reverts commit 1e502229e8484cd59584993e2ef5fa94d7bfd870.
15e2c9b3d9 Revert "Prioq can now remove any element if its position is known" This reverts commit e25b00cee82e185d0fe1447c64d07285a5fe5079.
8821c5989e Revert "Recursion to try a few different routes - score: 47.9M" This reverts commit 8187c0f1f24c5ee90fee56c8c42a88e404ba6059.
View comparison for these 19 commits »

8 år sedan

5ef8a78866 11M on input D by changing scoring function
72e8322f2b Revert "Some inputs perform better picking one of the next best rides" This reverts commit 1e502229e8484cd59584993e2ef5fa94d7bfd870.
abcac1b6fd Revert "Prioq can now remove any element if its position is known" This reverts commit e25b00cee82e185d0fe1447c64d07285a5fe5079.
654a9971c1 Revert "Recursion to try a few different routes - score: 47.9M" This reverts commit 8187c0f1f24c5ee90fee56c8c42a88e404ba6059.
d7f9d3c460 Revert "Check recursive algorithm gives a correct result on scoreboard" This reverts commit b1a7f7bdbf44ba606a2572a7709f632fe0bb9211.
View comparison for these 12 commits »

8 år sedan

8187c0f1f2 Recursion to try a few different routes - score: 47.9M Probably not a good strategy for three reasons : a) code is quite a bit more complex b) gain is all relative (only 15k on third input) and c) it's long to execute, already ~3.5 minutes Only score is calculated, not the actual resulting out files (it would require saving alongside bestTotalScore)
e25b00cee8 Prioq can now remove any element if its position is known
View comparison for these 2 commits »

8 år sedan

1e502229e8 Some inputs perform better picking one of the next best rides instead of picking the next best ride according to our score function With the few performing better, total score is now past 48M.
46ec5ca971 Some inputs perform better picking one of the next best rides instead of picking the next best ride according to our score function
View comparison for these 2 commits »

8 år sedan

46ec5ca971 Some inputs perform better picking one of the next best rides instead of picking the next best ride according to our score function
e6528c2394 At 48M picking one of the next best rides (instead of the best)
View comparison for these 2 commits »

8 år sedan