High-Performance VNS for the Max-Cut Problem using Commodity Graphics Hardware

The Max-Cut problem consists of finding a partition of the graph nodes into two subsets, such that the sum of the edge weights having endpoints in different subsets is maximized. This NPhard problem for non planar graphs has different applications in areas such as VLSI and ASIC design. In the last decade, consumer graphics cards (GPUs) have increased their power due to the computer games industry. These cards are now programmable and capable of processing huge amounts of data in a Streaming Pipelined Architecture. This paper proposes a high-performance GPU implementation of Variable Neighborhood Search (VNS) for the Max-Cut problem. This algorithm is tested and compared with the non-parallel implementation of VNS in CPU.