Objective variables are defined to construct an objective function. The objective function is a summation of all variables that are designated as objective-type. Variables are defined as objective function contributions by starting with **obj**. Thus, the variables *obj1*, *objective*, *object[1]* would be included in the objective function summation.

Additionally, slack variables are included in the objective function. These variables begin with the key letters **slk** and are defined with a lower bound of zero.

The objective function is always minimized with APMonitor. Maximizing an objective function is accomplished by minimizing the negative of the original objective.

! original objective maximize z

The objective is modified by minimizing the negative of the original objective function.

! modified objective minimize -z

The APMonitor objective is reported in the minimized form. Thus, a maximized objective with a result of +17 is reported as -17.

The objective function is retrieved as the parameter *nlc.objfcnval* in a programming script from the *apm_tag* function such as:

obj = apm_tag(s,a,'nlc.objfcnval')

This obtains the current objective function value with input arguments *s=server* and *a=application name*. An objective may consist of multiple objectives that are maximized or minimized. They are all converted to minimization functions and added together.

! Example model with an objective variable Parameters p1 = 5 Variables objective v1 > 6 Equations objective = (v1 - p1)^2 |

! Equivalent model with a minimize objective statement Parameters p1 = 5 Variables v1 > 6 Equations minimize (v1 - p1)^2 |

! Equivalent model with a maximize objective statement Parameters p1 = 5 Variables v1 > 6 Equations maximize -(v1 - p1)^2 |

Solution p1 = 5 v1 = 6 objective = 1 |

Retrieved from http://www.apmonitor.com/wiki/index.php/Main/ObjectiveVariables

Page last modified on April 02, 2017, at 05:50 PM