The Correlated Knowledge Gradient for Maximizing Expensive Continuous Functions with Noisy Observations using Gaussian Process Regression
Warren R. Scott, Peter I. Frazier, Warren B. Powell · 2010
We extend the concept of the correlated knowledge-gradient policy for ranking and selection to the case of continuous decision variables. We propose an approximate knowledge gradient for problems with continuous decision variables in the context of a Gaussian process regression model, along with an algorithm to maximize it. In the problem class considered, we use the approximate knowledge gradient to sequentially choose where to sample an expensive noisy function in order to find the maximum quickly.