Verfahren mit PaToH testen

[poettig]

Könnte problematisch sein, da PaToH anscheinend falsche Ergebnisse berechnet:

I again played around with PaToH and used their edge balancing feature [2]. Unfortunately both solution quality AND the reported block sizes (in terms of
hyperedges per block) seems to be wrong. I double checked the PaToH result both with my tool and with hMetis (and both tools report the same results, which
are significantly different from the results PaToH reports for its partition).
I'm on vacation tomorrow, but I'll contact the main author/developer of PaToH on Monday and ask whether I'm doing something wrong or edge-balancing
is actually broken in PaToH.

Falls nichts kommt, mal bei Sebastian nachfragen!

[lukashuebner]

Duplika von "2-Phasen Ansatz"