Experiments demonstrate that randomized SDP rounding achieves the Goemans-Williamson 0.878 ratio for both max-cut and weighted fractional cut-covering simultaneously after only about 128 ln m samples, with LP post-processing yielding better certificates than the theoretical algorithm.
The unique games conjecture, integrability gap for cut problems and embeddability of negative-type metrics intoℓ1
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Maximum Cuts and Fractional Cut Covers: A Computational Study of a Randomized Semidefinite Programming Approach
Experiments demonstrate that randomized SDP rounding achieves the Goemans-Williamson 0.878 ratio for both max-cut and weighted fractional cut-covering simultaneously after only about 128 ln m samples, with LP post-processing yielding better certificates than the theoretical algorithm.