Once a model is built, you can tell Parity Builder to solve for a suitable set of assignments. The default settings will generally work well, at least to get started, but you might want to consider tweaking a few of them. Click on the Settings tab to find controls for adjusting the settings. Note that when you save a project, changes you made to the settings are not saved.
Solution time limit
The most important setting is the time limit. The time limit control has three spinners, for hours, minutes and seconds. Set the limit to a time span with which you are comfortable. Do not worry about setting it to too short an interval; you will have the option to tell Parity Builder to resume solution if it reaches the time limit and you are not happy with the result.
Random generator seed
The solution method used by Parity Builder makes many decisions based on "random" numbers coming from a "random number generator". The numbers depend on a seed value, which you can set here. Any positive integer up to 9,999,999 is acceptable. Typically you can leave this alone, but if you want to try forming rosters for the same group of individuals a second time (hoping for better results), you might try a new seed.
Minimum solver passes
Parity Builder generates a random solution as a starting point, works to improve it, and then at some point gives up on improving the current solution and generates a new random starting point. This parameter dictates the minimum number of times Parity Builder will try to improve a solution before moving on to a new starting point. A good value for this parameter is impossible to predict. If you think Parity Builder is struggling with you problem, you might bump the minimum pass limit up or down.
Minimum improvement
Once Parity Builder has exceed the minimum pass limit described above, it will abandon attempts to improve a solution and restart with a new solution if it cannot improve the penalty of the current solution by at least this much. Again, you typically do not need to adjust this parameter; if your problem seems particularly difficult, you might try increasing or decreasing the minimum improvement limit.
For purposes of the tutorial, you can leave the default settings in place or modify them freely. You are now ready to solve the problem.
| previous: Build the model | next: Solve the model |
