Calendar
LightTools Training
May 21-25, 2012
Light Tec and OEC
Brussels, Belgium
Visit Light Tec website
LightTools Training
July 23-27, 2012
Shanghai OE Technology Co. Ltd.
Beijing, China
SOET website
LightTools Advanced Training
September 18-20, 2012
Cybernet Systems Taiwan
Hsinchu, Taiwan
Cybernet Systems Taiwan website
Strategies in Light Europe
September 18-20, 2012
Light Tec and OEC
Munich, Germany
SIL website
Photonex
October 17-18, 2012
Light Tec
Coventry, UK
Photonex website
Introduction to Illumination Design Using LightTools
October 22-25, 2012
Pasadena, California
More details
COM Macro Programming in LightTools
October 26, 2012
Pasadena, California
More details
For a complete list of LightTools events worldwide: visit our website
|
|
In this issue:
Choosing effective variables using the LightTools Parameter Sensitivity Utility
The effectiveness of any optimization depends on the availability of variables that can be changed to achieve a desired result. For the optimizer to converge to a solution quickly and efficiently, you should use only variables that have a substantial effect on the merit function. In LightTools, you can use the Parameter Sensitivity Utility (PSU) to help you learn which variables are effective variables in your model.
The PSU uses the same variables and merit functions that the optimizer does, so it is easy to run the PSU with your existing optimization setup.
- Start the Parameter Sensitivity Utility from the Optimization>Parameter Sensitivity menu.
- Select the variables and merit functions you want to assess in the Main window, and press the “>” button to add them to the study.
 Run the study by pressing the green arrow icon in the toolbar at the top of the window.
The PSU will scan through each parameter combination and assess the performance using your open LightTools model. Once the parameter study has completed, you can review the Results tab to learn how the merit function changes with each variable tested.
With these results, you can determine which variables are effective and which are not (i.e., “ineffective variables”). Effective variables will show a large slope to the merit function over the variable range. You can plot a 2D surface chart to compare the effects of two variables at once using the Plot 2D Results Using LumViewer controls in the lower right section of the Results window. For example, the following chart shows changes in the Y direction, but negligible change in the X dimension. This indicates that the Front Curvature variable is an effective variable, while the Lens Thickness variable is not and would be wise to exclude during an optimization.
The PSU is a good tool to use to “learn” about your model and what the important variables are to the success of your design. Incidentally, the PSU is also a good tool to explore about your merit function space and search for a good set of start parameters to launch the optimizer from. For more details about using the PSU, please see the LightTools Optimization User’s Guide. There are more general tips about optimization in the “Topics in LightTools Optimization” presentation on the Customer Support Portal.
New Optimization Video
LightTools features two optimization engines. In this 3-minute video, learn how to switch optimization engines and when you should consider using a different engine.
|
