Configuring the Pointer-2 Strategy Options

If you are using the Pointer-2 strategy, you can control options such as the number of sub-flow runs, design space type, and penalty value.

  1. Double-click the Exploration component icon .

    The Exploration Component Editor appears.

  2. From the Exploration Component Editor Exploration Strategy list, select Pointer-2.

  3. In the Pointer-2 Strategy Options area, enter or select the following:

    Option Description
    Allowed number of sub-flow runs This value sets the limit on the number of sub-flow runs that the strategy will perform.
    Subflow Batch Size (Parallel only) Pointer-2 attempts to submit design points in batches equal to the subflow batch size. This setting allows you to match your optimization setup to your available compute resources. For example, if you have 16 SEE Stations available to solve an optimization problem, you should turn on parallel and set the subflow batch size to 16 to keep all 16 stations working on the optimization problem at the same time. This option applies only when you have selected Execute in parallel. The default value is 4
    Automatic This option forces the Pointer-2 strategy to use a hybrid combination of all the optimization techniques.
    User Defined This option allows you to choose the optimization technique that the Pointer-2 strategy uses to solve the design optimization problem. Use the Optimizer selection list to select one of the following techniques:
    • Evol
    • Hooke-Jeeves
    • Sequential Quadratic Programming (NLPQL)
    • Nelder and Mead Downhill Simplex
    • Multifunction Optimization System Tool (MOST)
    • Multi-Objective Particle Swarm Algorithm
    (Automatic only) Design space topography type This option describes the general behavior of the output functions. Select one of the following:
    • smooth for output functions with no irregular or abrupt changes in the output values.
    • rough for output functions with numerical noise and/or irregular or abrupt changes in the output values.
    • discontinuous if your feasible domain has several disjoined regions.
    • unknown if you do not know any specific information about your output functions.
    Penalty value for failed runs This parameter represents the value of the Penalty parameter that is used for all failed subflow runs. The default value is 1.0E30.
    Objective value for failed runs This parameter represents the value of the Objective parameter that is used for all failed subflow runs. The default value is 1.0E30.
    Maximum Failed Runs If this option is selected, the specified value is used to set the maximum number of failed subflow evaluations that can be tolerated by the Pointer-2 strategy. If the number of failed runs exceeds this value, the Pointer-2 strategy will terminate execution. When this option is not selected, the Pointer-2 execution will continue despite any number of failed subflow runs.
    Use fixed random seed If this option is selected, the random number generator used by the optimization algorithm is seeded using the value specified in the corresponding text box. All executions of the Optimization component will use exactly the same sequence of random numbers and, therefore, will produce exactly the same design points. This arrangement is useful for debugging the optimization process when it is necessary to reproduce the same sequence of design points.

    If this option is not selected, the random number generator is seeded by using the clock time at the moment of execution.

    Execute in parallel Select this option to execute design points in parallel, small batches. The size of the batch depends on the technique that the Pointer-2 strategy is currently executing. It is related to the number of selected design variables for numeric techniques or to the size of the population for exploratory techniques.
    Restore optimum design point after execution Select this option to set parameters to the best design point after execution. If this option is not selected, the optimization keeps all parameters at their last values (last execution of the subflow) after execution.
    Re-execute optimum design point Select this option to force the Pointer-2 strategy to re-execute the optimum design point after the optimization ends. Re-execution is recommended if there are some output parameters that are not directly involved in optimization, but their values depend on the values of design variables (e.g., output file parameters).
    Use automatic variable scaling Select this option to scale design variables using their respective ranges (i.e., distances between their upper and lower bounds). Isight removes the Scale Factor column from the Variables tab because the values of scale factors are calculated automatically.

    Note: You must define all upper and lower bounds (or allowed values) for all design variables when this option is selected. If some design variables do not have bounds, you will not be able to use this option.

  4. Click OK to save your changes and to close the Exploration Component Editor.