CAGE User's Guide

Automated Tradeoff

You can use automated tradeoff to run an optimization routine and fill your tradeoff tables. Once you have set up an optimization you can run an automated tradeoff. As with any other tradeoff you need at least one table. You can apply an optimization to a cell or region of a tradeoff table and the tradeoff values found are used to fill the selected cells. You can then fill the entire table by extrapolation. You can examine the optimization results in more detail at the Optimization Output node in the usual way.

There is an example automated tradeoff in the tutorial chapter, Tutorial: Optimization and Automated Tradeoff.

Using Automated Tradeoff

1. You need a CAGE session with some models and a tradeoff containing some tables.
• See Tradeoff Calibrations for instructions on setting up a tradeoff. You could use the tradeoff tutorial to generate a suitable example session.

1. You also need to set up an optimization before you can run an automated tradeoff. Free and fixed variables, objectives, and constraints must be set up.
   • For an example work through the step-by-step tutorial to set up some optimizations and then apply them to a tradeoff table. See Tutorial: Optimization and Automated Tradeoff.

2. Go to the tradeoff table you want to automate. You must select some table cells to apply the optimization to. Select individual cells, or click and drag to select a rectangle of cells. The selected cells do not have to be adjacent. Note that if you define a large region with many cells it can take a long time to calculate a multiobjective optimization for each cell. Try a small region (say up to six cells) to begin with. Right-click selected cells and select Define Region or use the toolbar button.
3. In the tradeoff view, click a cell in a region.
4. To apply optimization: select Table -> Automated Tradeoff.
• The toolbox checks to see if the selected cell is part of a defined region. If not, an error dialog informs you that you can only perform automated tradeoff on cells in a defined region.
• If part of a region, then a dialog appears that allows an appropriate (defined below) optimization to be selected from the current project.

Note    You must set up an optimization to run before you can perform an automated tradeoff. You do this in the Optimization view. See also Setting Up Optimizations.

• The cell/region is made into a data set of operating points that is linked to be the primary operating point set in the optimization. If the optimization object does not already have a primary operating set defined, then a new one is created. Note that any existing operating point set is reset to the previous state at the end of the automated tradeoff.

5. The optimization is run as if you were clicking Run from the Optimization node. See Running Optimizations. The dialogs for the free variable ranges/initial values and any other variable fixed values not specified by the operating point set appear and take the values from the data dictionary as usual. Click OK when you are satisfied with the values in these, and the optimization runs.

• Results are placed in the tradeoff object, that is, values for the tables involving the free variables or values for the tables for constraint or objective models. If the routine applied gives more than one solution, for example, an NBI optimization, then only the values from the first solution are placed in the tradeoff tables. Every cell in the defined region is filled.

6. The cells of the region become part of the extrapolation mask (as if apply point has been applied); so if you want you can then click Extrapolate in the toolbar to fill the rest of the table from your optimized automated tradeoff.

• The output from the optimization appears in the usual way (as an Output node under the optimization object). As for any other optimization you can use the Optimization Output node views to investigate the output. See Optimization Output Node.

  

Selected Solution View

What Are Appropriate Optimizations?