Rockwell Automation Arena OptQuest Users Guide - Manual (Page 29)

25
4
4
• Setti
ng
Up
an
d Op
timiz
in
g
a M
ode
l
Setting Up and Optimizing a Model
Setting up and optimizing a model using OptQuest requires the following steps:
1. Create an Arena model of the problem.
2. Prepare your Arena model for optimization.
3. Start OptQuest and open an OptQuest (.opt) file.
4. Set up the optimization:
‡
Select the controls to optimize.
‡
Identify the responses to use in the objective and constraint expressions
‡
Specify any constraints.
‡
Specify the objective.
‡
Select optimization options.
5. Run the optimization.
6. Interpret the results.
7. Refine the solutions.
You perform steps 1 and 2 in Arena, 3 to 7 in OptQuest, and 6 in both.
Preparing the Arena model
Before using OptQuest, you must first develop an appropriate Arena model for your
problem. This entails building a well-tested simulation model and then defining the
controls and responses that you plan to use in your optimization model. You should refine
the Arena model and run several simulations to ensure that the model is working correctly
and that the results are what you expect.
After you define the control variables and response statistics in Arena, you can begin the
optimization process in OptQuest. The first step of this process is selecting controls to
optimize. The values of these controls will change with each simulation until OptQuest
finds values that yield the best value for the objective. For some analyses, you might fix
the values of certain controls and optimize the rest.
Controls
Variables or resources in your Arena model are called controls. Keep in mind that
OptQuest will provide values for selected controls to Arena. If the Arena model were to
override any of the control values that OptQuest provides during the simulation, it would
interfere with the optimization. Therefore, any automation or control logic in the Arena
model must be properly set up to work with OptQuest.
For example, control logic could be used to increase gradually the value of a variable, as
this may be an accurate representation of the system being modeled. In this situation,
OptQuest may be set to provide the beginning value of the variable, and then the control