search

MilaNedic / moarchiving

This repository contains the description and corresponding files for the "Computing and updating Hypervolume" project
MIT License
0 stars 0 forks source link

Hypervolume in 4D #7

Closed naceee closed 1 week ago

naceee commented 3 weeks ago

Original C code and our python implementation don't always produce the same result (but sometimes they do):

The following points we get a hypervolume of 0.37037902191204, while in the original implementation it is 0.371995494511654.

points = [
    [0.6394267984578837, 0.025010755222666936, 0.27502931836911926, 0.22321073814882275],
    [0.7364712141640124, 0.6766994874229113, 0.8921795677048454, 0.08693883262941615],
    [0.4219218196852704, 0.029797219438070344, 0.21863797480360336, 0.5053552881033624],
    [0.026535969683863625, 0.1988376506866485, 0.6498844377795232, 0.5449414806032167],
    [0.2204406220406967, 0.5892656838759087, 0.8094304566778266, 0.006498759678061017],
    [0.8058192518328079, 0.6981393949882269, 0.3402505165179919, 0.15547949981178155],
    [0.9572130722067812, 0.33659454511262676, 0.09274584338014791, 0.09671637683346401],
    [0.8474943663474598, 0.6037260313668911, 0.8071282732743802, 0.7297317866938179],
    [0.5362280914547007, 0.9731157639793706, 0.3785343772083535, 0.552040631273227],
    [0.8294046642529949, 0.6185197523642461, 0.8617069003107772, 0.577352145256762]
]

In this example with 100 points, the calculated hypervolumes differ even more (our: 0.6664533136930482, original: 0.790316084994994).

points_4d_100.txt

naceee commented 2 weeks ago

It looks like the problem is that the C code doesn't handle dominated points correctly. Running it only with the non-dominated poins for the above example, computes the same hypervolume as we do (0.37037902191204). 

nondom_points = [
    [0.6394267984578837, 0.025010755222666936, 0.27502931836911926, 0.22321073814882275],
    [0.4219218196852704, 0.029797219438070344, 0.21863797480360336, 0.5053552881033624],
    [0.026535969683863625, 0.1988376506866485, 0.6498844377795232, 0.5449414806032167],
    [0.2204406220406967, 0.5892656838759087, 0.8094304566778266, 0.006498759678061017],
    [0.8058192518328079, 0.6981393949882269, 0.3402505165179919, 0.15547949981178155],
    [0.9572130722067812, 0.33659454511262676, 0.09274584338014791, 0.09671637683346401]
]

I also tried running the other algorithm (-R flag) for computing hypervolume in 4D (they propose two algorithms), which confirms the results we obtained and computes the same hypervolume on both sets.

all points non-dominated points
C implementation 0.371995494511654 0.37037902191204
C implementation with flag -R 0.37037902191204 0.37037902191204
our implementation 0.37037902191204 0.37037902191204

I will check the paper if I find some more information about this (or maybe @MilaNedic remembers something about it?) - maybe it's a feature and not a bug :). Otherwise it would probably be good to make an issue on their github, just to point it out, even though it hasn't been active for a long time now.

naceee commented 1 week ago

I double checked that everything is working fine with our algorithm and created an issue on the Andreia's repostory, so I'm closing this issue here.