conatus.expert.max_growth

conatus.expert.max_growth(ss, t_bounds, *, xtol, n_scan=64)

(G_max, t_max) over t in t_bounds: scan of n_scan points (declared default MAX_GROWTH_N_SCAN = 64), then bounded Brent to xtol (required) around the best scan point. The answer is the best of the two interval ends, the best scan point and the refined candidate (finding 15: a maximum at t = 0, G = 1, was missed because only the Brent result was returned); an optimiser failure raises.

Parameters:

• ss (StateSpace)

• t_bounds (tuple[float, float])

• xtol (float)

• n_scan (int)

Return type:

tuple[float, float]