Abstract:
The idea of this paper stems to modify "the fact that most used optimization methods use a local quadratic representation of the objective function". It also arises from the fact that the objective function may not be represented perfectly by local quadratic representation functions and the global minimizer may be obtained for the general objective functions. Consequently, a new non-quadratic model algorithm, in this paper, is suggested for solving unconstrained nonlinear optimization problems, which modifies the classical conjugate gradient (CG) algorithms. The new algorithm is derived and evaluated numerically for some standard nonlinear test functions. The results indicate that in general the new algorithm has an improvement percentages on the previous some selected CG-algorithms
|