Once you have specified the objective function, constraints and optimization variables, the optimization can be run. To run the optimization, simply click the Optimize! button at the bottom of the Optimization dialog.
Note: During the optimization
process, each run is treated as a deterministic
simulation. You cannot run an optimization if your Monte Carlo
settings are specified to run multiple realizations. The settings must be
specified to either run a deterministic simulation or a single probabilistic simulation (in which case it will automatically
be treated as a deterministic simulation).
Clicking the Optimize! button will immediately take you to the other tab in the optimization dialog (Run Optimization) and begin the optimization. The Run Optimization tab tracks the progress of the optimization:
The Run Optimization tab is divided into two sections:
Best Function Value vs. Iterations: This section of the dialog provides a graph that plots the most optimal value of the objective function (on the y-axis) versus the number of iterations that have been completed (on the x-axis).
Top Results: This section displays the objective function and optimization variable values from the ten most optimal iterations.
This tab is continuously updated during the course of an optimization.
While the optimization is running, a Pause button will be available. Pressing it will temporarily stop the optimization (and switch the Pause button into a Resume button, which can be clicked to restart the simulation). There is also an Abort button, which immediately ends the optimization.
After the optimization is completed, the Pause and Abort buttons will switch back to Optimize! and Close buttons.
When the optimization completes, a message will be displayed indicating that the optimization has completed. The first row in the Top Ten Results section will display the optimal values of the optimization variables and the optimized value of the objective function. When you press Close to exit the optimization, you will be presented with a dialog with options for saving these results.
In some cases, the optimization may not converge. In this case, GoldSim will provide a warning message.
Some types of models may have multiple local optima. In such a case, GoldSim may not necessarily find the global optimum. There are, however, some steps you can take to increase the likelihood of finding the global optimum.
Warning: Warning messages
are not displayed when running an optimization (although fatal error messages
are).
Warning: When running
an optimization, Interrupt messages are never displayed. Moreover, unless
the Specify Action When Message is off or cannot be shown field in the
Interrupt is set to "Continue" or "Skip remainder of current Realization and Continue" (or the
Interrupt is disabled), any triggered Interrupts will result in a fatal error.
Learn more about:
Finding a Global Optimum in Complex Models with Multiple Optima
Understanding Optimization Warning Messages