Hierarchical Bayesian optimization algorithm

Abstract

The hierarchical Bayesian optimization algorithm (hBOA) solves nearly decomposable and hierarchical optimization problems scalably by combining concepts from evolutionary computation, machine learning and statistics. Since many complex real-world systems are nearly decomposable and hierarchical, hBOA is expected to provide scalable solutions for many complex real-world problems. This chapter describes hBOA and its predecessor, the Bayesian optimization algorithm (BOA), and outlines some of the most important theoretical and empirical results in this line of research. © Springer-Verlag Berlin Heidelberg 2006.