In this section, we provide details of the four modules of this invention. However, any one module of this invention can’t solve our problem alone. It is the interaction of all parts that allow the user to understand and improve the optimization models.
We begin by defining the optimization problem. Our objective is to maximize or minimize the sum of objective terms by selecting optimal feasible levels of the independent variables. The set of process limits or constraints define the feasible region. For example, upper and lower limits on the independent variables are constraints on the problems. Additionally, there are dependent variables that aid in defining the objective terms and/or constraints. These dependent variables are functions …show more content…
Generate optimal parameter and bound sensitivities to identify areas of model improvement
J.D. Terry, Ken Tyner, Bill Docter, John Righi
In sensitivity analysis, we determine the optimal sensitivity to parameter changes. These parameters may include prices, process limits, variable bounds, and other problem constants. While shadow values directly quantify the total objective incentive to constraint changes; this invention also analyzes incremental improvements in each objective term and other process variables.
In both linear and nonlinear optimization analyses, skilled users may perform sensitivity analysis for individual parameters (see Fiacco 1976; Ganesh and Biegler 1987). This invention improves this by expanding the practice to the systematic sensitivity of all process limits. Furthermore, this invention obtains an advantage by leveraging clever side-by-side layouts to improve individual sensitivity analyses using cross comparisons for more well-rounded analyses.
By understanding how problem constants affect the final solution, the user is better equipped to improve the model and the system. These improvements become an intuitive consequence of this module. Moreover, these improvements and the cross comparisons of numerous potential improvements, cause additional insights into the system to be apparent to the
We begin by defining the optimization problem. Our objective is to maximize or minimize the sum of objective terms by selecting optimal feasible levels of the independent variables. The set of process limits or constraints define the feasible region. For example, upper and lower limits on the independent variables are constraints on the problems. Additionally, there are dependent variables that aid in defining the objective terms and/or constraints. These dependent variables are functions …show more content…
Generate optimal parameter and bound sensitivities to identify areas of model improvement
J.D. Terry, Ken Tyner, Bill Docter, John Righi
In sensitivity analysis, we determine the optimal sensitivity to parameter changes. These parameters may include prices, process limits, variable bounds, and other problem constants. While shadow values directly quantify the total objective incentive to constraint changes; this invention also analyzes incremental improvements in each objective term and other process variables.
In both linear and nonlinear optimization analyses, skilled users may perform sensitivity analysis for individual parameters (see Fiacco 1976; Ganesh and Biegler 1987). This invention improves this by expanding the practice to the systematic sensitivity of all process limits. Furthermore, this invention obtains an advantage by leveraging clever side-by-side layouts to improve individual sensitivity analyses using cross comparisons for more well-rounded analyses.
By understanding how problem constants affect the final solution, the user is better equipped to improve the model and the system. These improvements become an intuitive consequence of this module. Moreover, these improvements and the cross comparisons of numerous potential improvements, cause additional insights into the system to be apparent to the