The Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm is an iterative method for solving non-linear optimization problems without constraints (Fletcher, 1987). It belongs to the quasi-Newtonian methods, in which the Hessian matrix of the second derivative is not calculated. Instead, the Hesse matrix is approximated. Newton's method and BFGS methods do not guarantee convergence unless the function has a quadratic Taylor expansion near the optimum. However, BFGS may have acceptable performance even when optimization occurs that is not continuous (Curtis and Que, 2015).
BFGS optimization does not require parameter configuration