You can hit the software predicted IBU’s by adding only a portion of the water when you add the initial DME/LME.
This approach reads like the "partial boil with late extract additions" approach (or "wort a/b" in HtB, 4e). For many styles of beer, it is an excellent approach. But it does have come with trade-offs and constraints.
One of the trade-offs is that there will be limits to the amount of hop oils that a volume of wort can hold.
IIRC, this number is frequently stated to be around 100 IBUs (but I don't know the source). There appears to be similar numbers in the data in Basic Brewing Radios (BBR)'s "November 1, 2018 - IBUs vs Wort Gravity and Hop Stand Temps" episode.
Adding back half the water at the end of the boil will lower the amount of hop oils that make it to the FV by about half (so perhaps a limit of 50 IBUs?).
Back to BBR (again). Some of their recent hop sampler episodes have included lab measured IBUs. A recent episode (IIRC, April 25, 2019 - Hop Sampler - Galena, Denali, El Dorado) included some discussion of software estimated IBUs for some of those beers. Again, IIRC, the software estimated IBUs were close enough to be useful -- but not close enough to claim "I hit all my numbers".
For me, "all models are wrong, but some are useful" seems to be a nice way to summarize mathematically estimated IBUs.