Complexity of Local Search for Euclidean Clustering Problems

Bodo Manthey, Nils Morawietz, Jesse van Rhijn, Frank Sommer

Abstract: We show that the simplest local search heuristics for two natural Euclidean clustering problems are PLS-hard. First, we show that the Hartigan-Wong method, which is essentially the Flip heuristic, for k-Means clustering is PLS-hard, even when k = 2. Second, we show the same result for the Flip heuristic for Max Cut, even when the edge weights are given by the (squared) Euclidean distances between the points in some set 𝒳 ⊆ R^d; a problem which is equivalent to Min Sum 2-Clustering.