Custom Profiler Overview
The Custom Profiler enables you to optimize factor settings without graphical output. The Custom Profiler can be used for problems of any size. It is especially useful for large problems where the standard graphical profiler has too many graphs to visualize well.
The Custom Profiler report has many fields in common with other profilers. The Benchmark field holds a value, or benchmark, of the predicted response. You can compare new results to the benchmark value and update the value based on the current factor settings.
The Optimization report enables you to specify the formula to be optimized and specifications for the optimization.
Figure 7.2 Custom Profiler
Example of the Custom Profiler
This example uses data that demonstrates the flow of water through a borehole that is drilled from the ground surface through two aquifers. You want to optimize the predicted value.
1. Select Help > Sample Data Library > Design Experiment and open Borehole Latin Hypercube.jmp
2. Select Graph > Custom Profiler.
3. Select prediction formula and click Y, Prediction Formula.
Figure 7.3 Completed Custom Profiler Launch Window
4. Click OK.
Note that the Benchmark value is 70.83. The Current Y value is also 70.83. This value is the predicted response with all factors set to their mean values.
5. In the Custom Profiler Report, click Optimize.
Figure 7.4 Custom Profiler Report after Optimization
The optimization routine found an optimum predicted response at 311.17. In order to obtain the optimum of 311.17, all factors are set either at their minimum or maximum values. The optimum is greater than the initial Benchmark value of 70.83.
Launch the Custom Profiler Platform
The Custom Profiler can be accessed in the following ways:
• The Custom Profiler can be accessed directly from the Graph menu. When you access the Custom Profiler in this way, the Custom Profiler launch window appears. See “Profiler Launch Windows” in the “Introduction to Profilers” chapter for details.
• The Custom Profiler can be accessed as a red triangle menu option in many modeling platforms. See “Where to Find JMP Profilers” in the “Introduction to Profilers” chapter for details about the availability of the Custom Profiler in different platforms.
• The Custom Profiler can be accessed from the Model Comparisons platform. Select Profiler from the Model Comparisons red triangle menu. Then, select Custom Profiler from the Profiler red triangle menu.
• 
The Custom Profiler can be accessed from the Formula Depot platform. Select Profiler from the Formula Depot red triangle menu and select the models to be profiled. Then, select Custom Profiler from the Profiler red triangle menu.
The Custom Profiler can be accessed from the Formula Depot platform. Select Profiler from the Formula Depot red triangle menu and select the models to be profiled. Then, select Custom Profiler from the Profiler red triangle menu.The Custom Profiler Report
The initial Custom Profiler report shows settings and controls for the factors, responses, and optimization.
Factor Settings and Controls
Figure 7.5 Factor Settings and Controls
Factor
The list of model factors.
Current X
The current value of each factor. Click in a box to change the value of a factor. The slider controls can also be used to change factor settings.
Lock
Enables you to lock a factor so that it is fixed when the optimization is performed. You can change a locked factor using the slider or clicking in the box in the Current X column. The lock applies only to the optimization.
Nominal Column
Unlabeled column to the right of the Lock column that lists the current value of nominal factors.
Note: The Current X column for nominal factors displays a coded (numeric) value for the current nominal factor.
Lo Limit
The lower limit for each factor. Click in a box to change the value.
High Limit
The upper limit for each factor. Click in a box to change the value.
Response Settings and Controls
Figure 7.6 Response Settings and Controls
Response
The list of one or more responses.
Current Y
The predicted response based on the current X settings. This value updates as the factor settings are changed.
Lo Limit
Enables you to set a lower limit for your response.
High Limit
Enables you to set an upper limit for your response.
Benchmark
A saved predicted value of the response. Initially, this value is set to the predicted value when all responses are at their mean value.
Reset Benchmark
Updates the benchmark value to the current predicted value.
Optimization Settings and Controls
Figure 7.7 Optimization Settings and Controls
Formula
The formula to be optimized. When a single response is used, the expression is the response column name. When multiple responses are used, the expression is a sum of desirability functions. You can edit the objective expression.
Objective
The current value of the objective function. When a single response is used, the objective expression is the predicted response. When multiple responses are used, the objective expression is the desirability function. For more information about desirability functions, see “Desirability Profiling and Optimization” in the “Profiler” chapter.
Trips
The number of random starts in the optimization algorithm. Each trip restarts the algorithm at a different starting point. This guards against finding local solutions.
Max Cycles
The maximum number of cycles used in the optimization algorithm. Each cycle is single pass through the input parameters and optimizes each one individually.
Max Iter
The maximum number of optimization iterations per cycle for each input parameter.
Convergence Limit
The upper limit for the convergence criterion for the optimization algorithm. If the convergence criterion becomes less than this value, the algorithm stops.
Convergence Criterion
The value of the convergence criterion for the optimization algorithm.
Maximize
Enables you to choose to maximize or minimize the objective function.
Optimize
Starts the optimization algorithm.
Custom Profiler Platform Options
Factor Settings
Contains options identical to the Factor Settings submenu in the Prediction Profiler. See “Factor Settings” in the “Profiler” chapter.
Log Iterations
Creates a data table that contains iterations of the optimization algorithm. The data table appears after the Optimize button is clicked.
Alter Linear Constraints
Enables you to add, change, or delete linear constraints. The constraints are used in the Custom Profiler. See “Linear Constraints” in the “Introduction to Profilers” chapter.
Save Linear Constraints
Saves existing linear constraints as a data table script that is named Constraint. See “Linear Constraints” in the “Introduction to Profilers” chapter.
Simulator
Launches the Simulator. See Chapter 8, “Simulator”.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.