The Conjugate Gradient Method in Comparison with the Evolution Optimization for Solving Nonlinear Problems

The strategies for solving non-linear problems can be classified into two classes of deterministic and stochastic methods. One representative of each class, the Conjugate Gradient Method (CGM) and the Evolution Optimization (EVO), are compared with respect to storage requirements and convergence behavior. Both methods are used to determine the modeling parameters of a GaAs-MESFET from measured S-parameters. Although both algorithms lead to the same result it was found that the CGM is about 5 times faster than the EVO.