Scalability of using Restricted Boltzmann Machines for combinatorial optimization

Malte Probst, Franz Rothlauf, Jörn Grahl

Abstract: Highlights•We use a Restricted Boltzmann Machine in an EDA for combinatorial optimization.•The RBM models variable dependencies and decomposes complex problems into parts.•RBM–EDA is competitive or superior to state of the art EDAs for four standard benchmark problems.•The scalability is superior in terms of computation times.