Global optimization using Bayesian heuristic approach

Abstract

Traditional optimization evaluates its results by estimating the maximal deviation. The Bayesian approach (BA) can be regarded as an indirect approach using heuristics by assessing a prior distribution. Using BA on the randomized heuristics, the Bayesian heuristic approach (BHA), provides a natural and convenient method to include expert knowledge, and a more flexible optimization means. In this paper, we introduce the basic concepts of BHA, discuss the basic problems and process of using BHA in the continuous and discrete global optimization, respectively, and make some comments on the advantages and disadvantages of BHA.