Downhill Simplex Method

The Downhill Simplex Method is based on the original approach of Nelder Mead. The simplex method is a local and gradient free search method, which uses a simplex as basic structure. In each iteration the simplex is tested to be reflected on the hyperplane spanned by the best designs. If no improvement was found, the simplex is contracted until a convergence is reached. As convergence criterion the maximum simplex edge length and the difference of the objective function is checked.

In optiSLang, the simplex method is extended for constraint optimization by a penalty approach. It can handle solver noise and even failed designs. For a small number of design variables and constraint conditions the simplex method converges quite fast. For a larger number, ARSM is more suitable.

Further information about methods of multidisciplinary optimization used in optiSLang can be found here.

Initialization Options

To access the options shown in the following table, double-click the Simplex system on the Scenery pane and switch to the Simplex tab.

OptionDescription
Initialization
Start rangeStart range factor to calculate initial simplex.
Computational aspects
Minimum iterationsMinimum number of iterations before testing for convergence.
Maximum iterationsMaximum number of iterations.
Convergence test
ObjectiveTolerance for objective convergence test.
ParameterTolerance for parameter convergence test.

Additional Options

To access the options shown in the following table, in any tab, click Show additional options.

OptionDescription
Limit maximum in parallel

Controls the resource usage of nodes in the system.

When the check box is cleared (default), a value is chosen to ensure the best possible utilization of the child nodes. When the check box is selected, set the value manually to specify how many designs are sent to child nodes, limiting the maximum degree of parallelism for all children. Ansys recommends keeping the check box clear.

Auto-save behavior

Select one of the following options:

  • No auto-save

  • Actor execution finished

  • Every n-th finished design (then select the number of designs from the text field)

  • Iteration finished (for optimization, reliability)

The project, including the database, is auto-saved (depending on defined interval) after calculating this node/system (either when the calculation succeeds or fails).

By default, all parametric and algorithm systems have Every nth finished design 1 design(s) selected, all other nodes have No auto-save selected.