Rockwell Automation Arena OptQuest Users Guide - Manual (Page 41)

4 • S ETTING U P AND O PTIMIZING A M ODEL
37
•
•
•
• •
4
• Setti
ng
Up
an
d Op
timiz
in
g
a M
ode
l
The bottom of the optimization window displays a graph that plots the trajectory of the
search for the best solution. Splitter bars allow you to resize the parts of the window.
When the optimization has finished or is stopped by the user, the current value cells on all
grids are cleared.
When the optimization completes, a Best Solutions node is added to the tree and the right
hand side shows the top 25 solutions.
Best Solutions
When an optimization completes or is stopped by the user, the top 25 solutions are
displayed in the Best Solutions form. You can choose to display more solutions by
changing the number in the View Solutions group and clicking the Refresh button. Or you
may choose to display the detail of any one solution by checking the Select box and then
clicking the View button.
The first row in the Best Solutions grid will be the best solution, the second row the
second best, and so on. The simulation column identifies the simulation that generated
that solution. For example, if the first row shows simulation 105, the best solution was
found at the 105th simulation.
The status column indicates whether the solution was constraint-feasible or infeasible.
OptQuest always assigns a higher value to solutions that are constraint feasible so the grid
will always show feasible solutions before infeasible solutions, even though the objective
value for a constraint-infeasible solution may be better than a constraint-feasible solution.
If the user is running with a varying number of replications, a Confidence column is
displayed in the grid. The Confidence column indicates whether the confidence interval
defined on the Options form was met or not met. A replications column displays how
many replications were performed.
Objective values graph The graph at the bottom displays the rate at which the best objective
value has changed during the course of the search. This is shown as a
plot of the best objective values as a function of the number of
simulations completed. (For optimizations with more than 1000
simulations, not every point is plotted on the graph.) Infeasible solutions
(solutions that violate one or more constraints) are plotted as dashed red
lines. Feasible solutions are plotted as solid green lines.