`scipy.optimize._lbfgsb:setulb`

Signature

built-insetulb(m, x, l, u, nbd, f, g, factr, pgtol, wa, iwa, task, lsave, isave, dsave, maxls, ln_task)

Limited-memory Broyden-Fletcher-Goldfarb-Shanno with bounds.

Low-level interface to the L-BFGS-B optimization algorithm. Performs a single iteration step.

m : int: Number of limited-memory variable metric corrections.
x : ndarray, shape (n,): Current solution vector, float64. Modified in-place.
l : ndarray, shape (n,): Lower bounds on variables, float64.
u : ndarray, shape (n,): Upper bounds on variables, float64.
nbd : ndarray, shape (n,): Bound type for each variable, int32. 0: unbounded, 1: only lower, 2: both, 3: only upper.
f : float: Current objective function value.
g : ndarray, shape (n,): Gradient of objective function at current point, float64.
factr : float: Factor for convergence test on objective function.
pgtol : float: Projected gradient tolerance for convergence.
wa : ndarray: Workspace array for internal computations, float64.
iwa : ndarray: Integer workspace array for internal computations, int32.
task : ndarray: Task control array encoding algorithm state, int32. Modified in-place.
lsave : ndarray: Logical state array, int32. Modified in-place.
isave : ndarray: Integer state array, int32. Modified in-place.
dsave : ndarray: Real state array, float64. Modified in-place.
maxls : int: Maximum number of line search steps per iteration.
ln_task : ndarray: Task string length array, int32.