simulation multi_run set
Allows you to set the parameters for the multi-run simulation.
Format:
simulation multi_run set |
|---|
load_analyses = | yes/no |
save_analyses = | yes/no |
analysis_prefix = | string |
stop_on_error = | yes/no |
save_curves = | yes/no |
chart_objectives = | yes/no |
chart_variables = | yes/no |
show_summary = | yes/no |
opt_algorithm = | yes/no |
opt_maximum_iterations = | integer |
opt_convergence_tolerance = | real |
opt_differencing_technique = | centered_difference/forward_difference |
opt_scaled_perturbation = | real |
opt_user_parameters = | real |
opt_rescale_iterations = | integer |
opt_slp_convergence_iter = | integer |
opt_debug = | on/off |
Example:
simulation multi_run set & |
|---|
load_analysis = | yes & |
save_analyses = | yes & |
analysis_prefix = | ANAL_1 & |
stop_on_error = | yes & |
save_curves = | no & |
show_summary = | no & |
opt_user_parameters = | 2.0 |
Description:
Parameter | Value Type | Description |
|---|
load_analysis | Yes/No | Specifies yes or no. |
save_analysis | Yes/No | Set to Yes to automatically copies the parametric analysis results to a permanent location when the analysis is complete |
analysis_prefix | String | Enters the name you want to use for each analysis object. |
stop_on_error | Yes/No | Set to Yes to stop the parametric analysis if Adams Solver encounters an error during a simulation. If you set it to No, Adams Solver continues running simulations even if a simulation fails or another error occurs. |
save_curves | Yes/No | Clears all displayed measures at the beginning of the parametric analysis and automatically saves the curve from each trial or iteration. |
chart_ objectives | Yes/No | Enter yes or no. See extended definition for more details. |
chart_variables | Yes/No | Displays a strip chart for each design variable, plotting its value versus the trial or iteration number. Adams View updates the strip chart every trial or iteration. Set it to yes or no accordingly. |
show_summary | Yes/No |
opt_algorithm | Algorithm | Specify optimization algorithm. See extended definition for more details. |
opt_maximum_ iterations | Integer | Maximum iterations tells the optimization algorithm how many iterations it should take before it admits failure. Note that a single iteration can have an arbitrarily large number of analysis runs |
opt_ convergence_ tolerance | Real | Convergence tolerance is the limit below which subsequent differences of the objective must fall before an optimization is considered successful. |
opt_ differencing_ tecnique | Centered_difference/forward_difference | The differencing technique controls how the optimizer computes gradients for the design functions. |
opt_scaled_ perturbation | Integer | The size of the perturbation can reduce the effect of errors in the analysis. |
opt_user_parameter | Integer | Adams View passes the user parameters to a user-written optimization algorithm. |
opt_rescale_iteration | Real | Rescale iteration is the number of iterations after which the design variable values are rescaled. If you set the value to -1, scaling is turned off. |
opt_slp_convergence_iter | Real | The number of consecutive iterations for which the absolute or relative convergence criteria must be met to indicate convergence in the DOT Sequential Linear Programming method. |
opt_debug | On/Off | Turning on debugging output sends copious optimizer diagnostics to the window that launched Adams View. |
Extended Definition:
1. For the “analyis_prefix” parameter, Adams View appends a unique number to the prefix to form the complete name of the new analysis object. Adams View creates the new analysis under the model you analyzed.
2. If you do not select Save Curves, Adams View does not clear or save any curves. It only displays the curve for the current simulation and any curves you previously saved. Set to yes to save curves, else, set to No.
3. For the “Chart_ Objectives” parameter, enter “yes” to display a strip chart of the following, depending on the type of parametric analysis:
■Objective value versus variable value for a design study.
■Objective value versus trial for a design of experiments (DOE).
■Objective value versus iteration number for an optimization.
Adams View updates the strip chart at every trial or iteration.
4. Optimization Algorithm (opt_algorithm = Algorithm)
Algorithm specifies the algorithm used to perform the optimization. The MSCADS algorithms are provided with Adams View. The DOT algorithms can be purchased from Vanderplaats R&D, Inc. You can also include your own optimization algorithm.
■mscads_mmfd - Use the MMFD (Modified Method of Feasible Directions) algorithm from MSC Automated Design Synthesis code. This algorithm requires that design variables have range limits, since it works in scaled space.
■mscads_sqp - Use the SQP (Sequential Quadratic Programming) algorithm from MSC Automated Design Synthesis. This algorithm requires that design variables have range limits, since it works in scaled space.
■mscads_slp - Use the SLP (Sequential Linear Programming) algorithm from MSC Automated Design Synthesis. This algorithm requires that design variables have range limits, since it works in scaled space.
■mscads_sumt - Use the SUMT (Sequential Unconstrained Minimization Technique) algorithm from MSC Automated Design Synthesis. This algorithm requires that design variables have range limits, since it works in scaled space.
■mscads_opt - Enables a user selectable optimization strategy as documented in Vanderplaats, G.N., ADS - A Fortran program for Automated Design Synthesis - Version 1.10, NASA CR 177985, 1985, Appendix A. The default is recommended.
■optdes_grg - This option is still provided for legacy model support. When specified, Adams will use the MSCADS-MMFD option and issue a warning message.
■optdes_sqp - This option is still provided for legacy model support. When specified, Adams will use the MSCADS-SQP option and issue a warning message.
■dot1 - Use DOT with BFGS (Broydon-Fletcher-Goldfarb-Shanno) for unconstrained problems. Use DOT with MMFD (Modified Method of Feasible Directions) for constrained problems.
■dot2 - Use DOT with FR (Fletcher-Reeves) for unconstrained problems. Use DOT with SLP (Sequential Linear Programming) for constrained problems.
■dot3 - Use DOT with FR (same as DOT2) for unconstrained problems. Use DOT with SQP (Sequential Quadratic Programming) for constrained problems.
■user1, user2, user3 - Allows you to invoke a user-written optimization algorithm that has been linked to Adams View. See
Linking External Algorithms for more information on how to link an external optimization algorithm to Adams View and associate it to user1, user2, or user3.
5. For the “opt_convergence_tolerance” parameter, if the condition ABS (objective[now] - objective[now-1]) < convergence_tolerance is true for a certain number of iterations (usually two), then the convergence tolerance criterion is met. Note that this is only one test that is made by most optimization algorithms before they terminate successfully.
Like other Adams Solver tolerances, you may need to experiment with this tolerance to find the right value for your application. Display the objective versus iteration strip chart. If the optimizer quits even though the last iteration made noticeable progress, try reducing the tolerance. If the optimizer continues iterating even after the objective has stopped changing very much, make the tolerance larger.
6. Centered differencing perturbs each design variable in the negative direction from the nominal value, then again in the positive direction using finite differencing between the perturbed results to compute the gradient. If you choose forward differencing, each design variable is perturbed in a positive direction only.
Centered differencing can sometimes generate smoother, more reliable gradients (especially in noisy models), but it causes twice as many analysis runs to be performed.
7. Naturally, you want to remove as much of pertubations as you can. If you are uncertain of the accuracy and smoothness of your model, use a large perturbation at first, then reduce it as you get better designs. Remember that the accuracy of the gradients generally improves as the perturbations get smaller.
8. Realizing that there may be parameter information that is not conveyed through the existing parameter set, this “opt_user_parameter” parameter was added to allow you to pass any real numeric data to your algorithm.
9. Keep an eye on the debug window as some important warnings might be written there.The debugging output shows you the data the optimizer is receiving from Adams View, among other things. If the optimizer is behaving erratically, this may help you determine the source of the problem.
Cautions:
1. Use care if you turn the “stop_on_error” option off. Optimizations probably do not recover well from an error. In some cases, you may want to continue a design study or design of experiment even if a few of the simulations fail.