On the optimization of site investigation programs using centroidal Voronoi tessellation and random field theory

This work proposes a centroidal Voronoi tessellation (CVT)-based site investigation scheme, which is applicable to arbitrary site geometries and arbitrary numbers of investigation locations. A modified Lloyd algorithm is developed to generate CVT-based site investigation programs. With the data from prior site investigations, the random field statistics of a site property are estimated using a Markov chain Monte Carlo simulation-based Bayesian inference approach, and the site property at each location is estimated using Kriging interpolation. The performance of the CVT-based site investigation programs and optimization are demonstrated using two illustrative examples, namely, a square site and an irregular site.