Efficient Frontier Runtime Tab

Figure 1 - Optimization Settings - Efficient Frontier Runtime Tab

The Efficient Frontier Runtime tab displays Evolver settings that determine the runtime of an Efficient Frontier analysis.

If no stopping conditions are selected, Evolver will run until all possible solutions have been tried, or until the analysis is stopped manually by pressing the Stop button in the Progress or the Evolver Watcher windows.

When multiple conditions are checked, Evolver will stop as soon as any one of the chosen conditions is met. Stopping conditions can be manually overridden at any time by using the Stop button in the Evolver Watcher or Progress windows.

The options on the Efficient Frontier Runtime tab include:

  • Trials per Constraining Value - The number of trials that will be run per constraining value for the Efficient Frontier analysis. For example, if Trials is set to 10,000, and the Efficient Frontier analysis has 6 alternative values in a constraint, the Efficient Frontier analysis will stop after 60,000 trial solutions (6 x 10000 simulations).
  • Time - Stop Evolver from analyzing scenarios after the given number of hours, minutes, or seconds has elapsed.
  • Progress - Stop Evolver from generating scenarios when the improvement in the target cell value is less than the specified amount.
    • Maximum Change - The greatest degree of change that Evolver will recognize as valid and continue analysis. This option can be expressed as a percent or as a static value using the pulldown to the right of the value.
    • Number of Trials - The number of trials over which to check the improvement. The more complex the optimization problem, the larger the number of trials you should specify in this stopping condition.

    • This is the perhaps most popular stopping condition, as it provides an effective way to stop analysis after the improvement rate is slowing and better solutions are likely not possible.

      For example, when trying to maximize the value of the target cell, the values used in the Progress stopping condition could be 20000 trials, and 0.01% change. In this scenario, suppose that after 10,000 trials, the best target cell value found is 2000.  If the Progress option is the only stopping condition selected, when Evolver gets to trial 30,000 (20,000 successive trials), analysis will continue only if Evolver is able to find a target cell value of at least 2000.2 during those last 20000 trials (a 0.01% change to the best value found after 10,000 trials). In other words, if the target cell values have not improved at least 0.01% over the last 20,000 trials, Evolver concludes there is little more improvement to be found, and it stops the search.