A Modified q-BFGS Algorithm for Unconstrained Optimization

This paper presents a modification of the q-BFGS method for nonlinear unconstrained optimization problems. For this modification, we use simple symmetric positive definite matrix and propose new q-quasi-Newton equation, which is close to ordinary equation in limiting case. uses only first order q-derivatives build an approximate q-Hessian over number iterations. The q-Armijo-Wolfe line search condition used calculate step length, guarantees that objective function value decreasing. modified preserves global convergence properties method, without convexity assumption on function. Numerical results some test problems are presented, show improvement has been achieved. Moreover, depict numerical through performance profiles.