OptQuest

Improved Search with Design Variables

OptQuest supports design variables in a simulation. These are nonnumeric variables that can take one of multiple possible values and are used for decisions where the value of the variable represents an alternative, and not a quantity. An example would be a simulation that included a choice of schedules to implement with three choices: Schedule A, Schedule B, or Schedule C.

Algorithms have been updated in this release related to how design variables are handled during the optimization process that should lead to better solutions and more rapid convergence in problems involving design variables.

Improved Diversity Search

OptQuest cycles through diversity and intensification phases during the optimization search process. In a recent release a faster diversity algorithm was introduced that reduced OptQuest computational time. However, this algorithm exhibited some correlations between pairs of variables in higher dimension optimizations. These correlations did not negatively impact the performance of the OptQuest search, but in this release we updated the diversity algorithm today to still be very fast, but so that this correlation in higher dimensions does not happen.

The new diversity algorithm relies in part on new tables that are embedded in the code and the inclusion of these tables, especially to deal with the cases of optimization problems with a large number of variables has increased the size of the delivered OptQuest library.

Additional Information Captured in LogSetup()

OptQuest has a LogSetup() method that can be used to log to a file all of the problem information currently loaded in a COptQuestOptimization object and some information about the current state. This method is helpful to check the status of your model and is also used for troubleshooting issues when support is requested from OptTek.

In this release additional information was added to the information that is logged with this method call:

• Constraint Marked as Goal
• Stop Solution Generation
• Replication Batch Size
• Is Running Serial Replication
• Frontier Multi Objective Min and Max

Support for Handling Nonlinear Constraints of Input Variables as Goals

Handling nonlinear constraints of input variables prior to running the simulation was added in the last release, but marking these constraints as goals was not supported. The capability to mark nonlinear constraints of input variables as goals is now supported.

C# DLL Compiled with .NET Standard 2.0

The C# DLL version of OptQuest is now compiled with .NET Standard 2.0. Let us know if you need the C# DLL compiled against a different target.

Several minor bug fixes are also included in this release. Contact your OptQuest representative for more details or with specific questions.