ternary_search

mlquantify.solvers.ternary_search(left, right, objective, tol=1e-06)

Find the minimum of a unimodal function via ternary search.

Iteratively narrows the interval [left, right] by comparing the objective at two interior probe points until the interval width falls below tol.

Parameters:

leftfloat

Left bound of the search interval.

rightfloat

Right bound of the search interval.

objectivecallable

Scalar function to minimize over [left, right]. Must be unimodal on the interval.

tolfloat, default=1e-6

Convergence criterion: the search stops when right - left <= tol.

Returns:

minimumfloat

Approximate location of the minimum.

See also

solve_binary

Binary prevalence solver that can call this search.

Examples

>>> from mlquantify.solvers._binary import ternary_search
>>> minimum = ternary_search(0.0, 1.0, lambda x: (x - 0.3) ** 2)
>>> round(minimum, 4)
0.3