Tolerance and Priority Parameters

<< Click to Display Table of Contents >>

Navigation:  Chapter 3 - GeneHunter Excel Interface Reference >

Tolerance and Priority Parameters

Previous pageReturn to chapter overviewNext page

The user can add a tolerance and a priority for each soft constraint entered.  The purpose of the tolerance is to tell GeneHunter when it has done a sufficient job.  The purpose of the priority parameter is to tell GeneHunter which soft constraints are more important than others.


To specify soft constraints:

1.  In the show list options in the GeneHunter Dialog Screen , click on constraints.

2.  Click on the Add button to display the dialog box below.


Figure 3 Change soft constraints


3.  Either type in the cell(s) reference or select the cell(s) on the worksheet using the mouse.

4.  Click on the arrow and select whether you want the constraint cell to be greater than, less than, or equal to a value in the condition box.

5.  If you want the adjustable cell to be equal to a value in the condition box, either type that value in the condition edit box or select a cell or cells from the worksheet using the mouse.  Note that if you add more than one cell in the cell reference box, there must be an equal number of cells in the condition box.

6.  Set the Tolerance value by typing in a number in the edit box.  The purpose of the tolerance value is to tell GeneHunter when it has done a sufficient job in satisfying a soft constraint. For example, if you specify a constraint that cell B4 must be less than cell B6, then you may want to tell GeneHunter that if the values in the cells are within plus or minus 0.5, then the solution is acceptable. To accomplish this, type 0.5 in the Tolerance edit box.  Selecting the correct Tolerance value is based upon your knowledge of the problem.  Note:  GeneHunter 2.0 employs a special algorithm of tolerance processing in the case when tolerance is set to zero for constraints. It sets the tolerance internally to a large non-zero value just to allow some number of feasible solutions to emerge. These feasible solutions compete based on the fitness function value, as modified internally by the constraints. This algorithm provides a better opportunity for the population to generate feasible solutions. Otherwise, if the starting population does not contain enough feasible solutions, it has no realistic chance to produce them. In other words, this auto-tolerance algorithm helps to a great extent to optimize highly constrained models. During evolution the internal tolerance eventually decreases to a very small number. If you want to eliminate this auto-tolerance feature, just set the tolerance to a very small number instead of zero.

7.  Set the Priority for the constraint by clicking on the arrow and selecting either high, medium, or low with the mouse.  If there is more than one soft constraint, GeneHunter uses this value to satisfy the constraints in the order you specify.