HFSS15: Choosing an Optimizer

Conducting an optimization analysis allows you to determine an optimum solution for your problem. In HFSS optimization analyses, you have choices of optimizer, though in most cases, the Sequential Nonlinear Programming optimizer is recommended.

• Sequential Nonlinear Programming (SNLP)

• Sequential Mixed Integer NonLinear Programming (SMINLP)

• Quasi Newton

• Pattern Search

• Genetic Algorithm

• MATLAB

All optimizers assume that the nominal problem you are analyzing is close to the optimal solution; therefore, you must specify a domain that contains the region in which you expect to reach the optimum value.

All optimizers allow you to define a maximum limit to the number of iterations to be executed. This prevents you from consuming your remaining computing resources and allows you to analyze the obtained solutions. From this reduced range, you can further narrow the domain of the problem and regenerate the solutions.

All optimizers also allow you to enter a coefficient in the Add Constraints window to define the linear relationship between the selected variables and the entered constraint value. For the SNLP and SMINLP optimizers, the relationship can be linear or nonlinear. For the Quasi Newton and Pattern Search optimizers, the relationship must be linear.