The computation of one or more probability distributions or measures is required for a number of problems in statistics and data science. Optimal experimental design, or more particularly, optimal approximate design, is one example. Other examples include parameter estimation, adaptive design, stratified sampling and image processing.