Main

## TIER - APMonitor Option

## Main.OptionTier History

Hide minor edits - Show changes to output

Changed line 12 from:

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an MV to be a degree of freedom. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an CV to be included in the objective function. If there is a sub-optimization problem that has no MV degrees of freedom, a warning message is displayed. Consider the situation with 2 MVs and 2 CVs with TIER 1 and 2 for each type (~~MV1~~,~~MV2~~,~~CV1~~,~~CV2~~). The solver would first optimize the set ~~MV1~~ and ~~CV1~~. It would then turn off the ~~MV1~~ status and fix it at the optimized values and remove ~~CV1~~ objective function contribution. It would then optimize the set ~~MV2~~ and ~~CV2~~. The optimization occurs in two separate steps but the final reported solution includes the new ~~MV1~~ and ~~MV2~~ optimized profiles.

to:

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an MV to be a degree of freedom. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an CV to be included in the objective function. If there is a sub-optimization problem that has no MV degrees of freedom, a warning message is displayed. Consider the situation with 2 MVs and 2 CVs with TIER 1 and 2 for each type (MV'_1_',MV'_2_',CV'_1_',CV'_2_'). The solver would first optimize the set MV'_1_' and CV'_1_'. It would then turn off the MV'_1_' status and fix it at the optimized values and remove the CV'_1_' objective function contribution. It would then optimize the set MV'_2_' and CV'_2_'. The optimization occurs in two separate steps but the final reported solution includes the new MV'_1_' and MV'_2_' optimized profiles.

Changed line 12 from:

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an MV to be a degree of freedom. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an CV to be included in the objective function. If there is a sub-optimization problem that has no MV degrees of freedom, a warning message is displayed.

to:

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an MV to be a degree of freedom. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an CV to be included in the objective function. If there is a sub-optimization problem that has no MV degrees of freedom, a warning message is displayed. Consider the situation with 2 MVs and 2 CVs with TIER 1 and 2 for each type (MV1,MV2,CV1,CV2). The solver would first optimize the set MV1 and CV1. It would then turn off the MV1 status and fix it at the optimized values and remove CV1 objective function contribution. It would then optimize the set MV2 and CV2. The optimization occurs in two separate steps but the final reported solution includes the new MV1 and MV2 optimized profiles.

Changed line 5 from:

%width=50px%Attach:mv.png %width=50px%Attach:cv.png

to:

%width=50px%Attach:mv.png %width=50px%Attach:cv.png [[Main/DbsVariable|Local Options]] | %width=30px%Attach:apm.png [[Main/DbsGlobal|Global Options]]

Changed line 12 from:

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem.

to:

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an MV to be a degree of freedom. The [[Main/OptionStatus|STATUS]] must also be ON (1) for an CV to be included in the objective function. If there is a sub-optimization problem that has no MV degrees of freedom, a warning message is displayed.

Added lines 1-14:

(:title TIER - APMonitor Option:)

(:keywords TIER, Optimization, Model predictive control, Option, Configure, Default, Description:)

(:description Tiered order of Manipulated Variables (MVs) and Controlled Variables (CVs):)

%width=50px%Attach:mv.png %width=50px%Attach:cv.png

Type: Integer, Input

Default Value: 1

Description: Tiered order of Manipulated Variables (MVs)

and Controlled Variables (CVs)

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem.

See also [[https://apmonitor.com/do/index.php/Main/MultiObjectiveOptimization|Multi-Objective Optimization]]

(:keywords TIER, Optimization, Model predictive control, Option, Configure, Default, Description:)

(:description Tiered order of Manipulated Variables (MVs) and Controlled Variables (CVs):)

%width=50px%Attach:mv.png %width=50px%Attach:cv.png

Type: Integer, Input

Default Value: 1

Description: Tiered order of Manipulated Variables (MVs)

and Controlled Variables (CVs)

TIER is an ascending order of precedence for optimization. Tuning an application with multiple objectives can be challenging to coordinate the weights of competing objectives although there is a clear rank priority order. TIER gives the option to split the degrees of freedom into multiple sub-optimization problems. The highest priority values are optimized first while the subsequent priority values are optimized as a next step. Valid TIER values for MVs and CVs are between 1 and 100. There are up to 100 optimization levels and individual MVs / CVs must be at the same TIER value to be included in the sub-optimization problem.

See also [[https://apmonitor.com/do/index.php/Main/MultiObjectiveOptimization|Multi-Objective Optimization]]