A Local Search Approximation Algorithm for Max-k-Cut of Graph and Hypergraph

Abstract

Given a graph or hyper graph, the graph or hyper graph Max-k-Cut problem is to partition the vertices into k nonempty sets such that the sum of weights of edges across different sets is maximized. We present a deterministic local search algorithm for the problem, which has a performance ratio 1 - 1/k for Max-k-Cut of graph, and has a similar result for Max-k-Cut of hyper graph.