Figure 2:

An example of a policy lattice for five quasi-identifying values and an illustration of the main functions of the SHS algorithm. (a) depicts a random path initialization of the frontier composed of five policies (p₁, p₂, p₃, p₄, and p₅). (b) depicts a sublattice structure sublattice (p_a, p_b) in the policy space and its corresponding bounding region in the R-U space, the yellow area of which is not dominated by the frontier. The proportion of the yellow area is used as the heuristic of the probability that a randomly selected policy in the sublattice can update current frontier. (c) depicts the updates to the frontier: p_a, p_b, and p_i from the sublattice (p_a, p_b) will be on the frontier while policy p₄ is removed.