Rockwell Automation Arena OptQuest Users Guide - Manual (Page 28)

A RENA O PT Q UEST U SER ’ S G UIDE
24
• •
•
•
•
Types of optimization models
Optimization models can be classified according to the control types as:
An optimization model can also be classified according to the functional forms used to
define the objective and the constraints. Hence, an optimization model can be linear or
nonlinear. In a linear model, all terms in the formulas consist of a single control multiplied
by a constant. For example, 3*x – 1.2*y is a linear relationship since both the first and
second term only involve constants multiplied by controls (in this case, x and y).
Terms such as x 2 , x*y, or 1/x make nonlinear relationships. Any models that contain such
terms in either the objective or a constraint are classified as nonlinear.
A third classification casts optimization models as deterministic or stochastic (i.e., a
model or system with one or more random elements), depending on the nature of the
model data. In a deterministic model, all input data is constant or assumed to be known
with certainty. In a stochastic model, some of the model data is uncertain and is described
with probability distributions. Stochastic models are much more difficult to optimize
because they require simulation to compute the objective function. While OptQuest is
designed to solve stochastic models using Arena as the objective function evaluator, it is
also capable of solving deterministic models.
Model Control Type
Discrete Only discrete controls
Continuous Only continuous controls
Mixed Both discrete and continuous controls