One-parameter exercise. Subpopulations of selected planets (in orange) for the leverage objective function (6), using alternative heuristics for maximizing the leverage objective function: regular classes (Section 3.3.1, top left panel), quantile classes (Section 3.3.2, top right panel), K-means clustering (Section 3.3.3, bottom-left panel), and simulated annealing (Section 3.3.5, bottom right panel). The selection using the leverage-greedy algorithm of Section 3.3.4 was already shown in the bottom panel of Figure 7.