About Design Parameters Initialization

Each process component (DOE, Optimization, etc.) initializes their design parameters (e.g., factors, design variables, constraints, etc.) using information from the general problem formulation for that component. For example, factors selected in a DOE component will use any defined lower and upper bound in the general problem formulation to define the range over which to vary the factor in the set of experiments. Similarly, the Optimization component will use the lower and upper bound information from the problem formulation to define the space in which the optimization algorithm is allowed to vary design variables and to define the feasible range for selected outputs (i.e., constraints). The following sections explain how each individual process component initializes their design parameters using the information from the general problem formulation.

DOE Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following DOE component design parameter attributes:

Factors

The Lower and Upper bounds specified on the input parameters in the general problem formulation are used as Lower and Upper levels for the factors.

Note: The values defined in the general problem formulation are true values for the parameter and might be internally converted to level values depending on whether the Relation attribute for the factor is set to values, %, or diff.

Responses

The objectives specified on output parameters in the general problem formulation are used to initialize the Objective attribute for corresponding selected responses. If the objective is defined to be a target, the target value from the general problem formulation is used to initialize the Target attribute. The Objective Weight from the general problem formulation is used to initialize the Weight attribute of any DOE objective.

Loop Component

The Loop component does not support problem formulations.

Monte Carlo Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following Monte Carlo component design parameter attributes:

Random Variables

By nature, random variables are defined by some distribution around a mean value; therefore, bounds cannot be used to specify any information about the random variables in a meaningful way.

Responses

The Lower and Upper bounds specified on output parameters in the general problem formulation are used as Lower and Upper limits for reliability calculations for any defined responses.

Optimization Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following Optimization component design parameter attributes:

Design Variables

The Lower and Upper bounds specified on input parameters in the general problem formulation are used as Lower and Upper bounds for design variables.

Constraints

The Lower and Upper bounds specified on output parameters in the general problem formulation are used as Lower and Upper bounds for constraints. The Bound Scale and Bound Weight from the general problem formulation are used to initialize the Scale and Weight attributes of any defined constraints.

Objectives

Any parameter that has an objective defined in the general problem formulation will be initialized automatically as an objective with the objective direction (minimize/maximize/target) as specified in the general problem formulation. If the parameter is an input parameter, it will also be initialized as a design variable because optimization requires that it be able to be varied if it is an objective. The Objective Scale and Objective Weight from the general problem formulation are used to initialize the Scale and Weight attributes of any defined objective.

SDI Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following SDI component design parameter attributes:

Design Variables

The Lower and Upper bounds specified on input parameters in the general problem formulation are used as Lower and Upper bounds for design variables.

Random Variables

By nature, random variables are defined by some distribution around a mean value; therefore, bounds cannot be used to specify any information about the random variables in a meaningful way.

Responses

The Lower and Upper bounds specified on output parameters in the general problem formulation are used as Lower and Upper limits for reliability calculations for any defined responses.

Six Sigma Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following Six Sigma component design parameter attributes:

Random Variables

By nature, random variables are defined by some distribution around a mean value; therefore, bounds cannot be used to specify any information about the random variables in a meaningful way.

Responses

The Lower and Upper bounds specified on output parameters in the general problem formulation are used as Lower and Upper limits for reliability calculations for any defined responses.

Factors

The Lower and Upper bounds specified on input parameters in the general problem formulation are used as Lower and Upper levels for defined factors.

Note: The values defined in the problem formulation are true values for the parameter and might be internally converted to level values depending on whether the Relation attribute for the factor is set to values, %, or diff.

Taguchi Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following Taguchi component design parameter attributes:

Control Factors

The Lower and Upper bounds specified on input parameters in the general problem formulation are used as Lower and Upper levels for Control factors. If the Control factor is also a Noise factor, the Lower and Upper bounds specified on the input parameters are adjusted to account for Noise factor relative perturbations when setting Control factor levels. For example, if the Lower and Upper bounds for parameter X are set to 9 and 22, and X is defined as a Control factor and a Noise factor, with Noise factor levels of +/- 10 %, the Control factor levels will be set to 10 and 20.

Signal Factors

The Lower and Upper bounds specified on input parameters in the general problem formulation are used as the Lower and Upper levels for Signal factors.

Note: The values defined in the problem formulation are true values for the parameter and might be internally converted to level values depending on whether the Relation attribute for the factor is set to values, %, or diff.

Target Solver Component

The general problem formulation that you define (from the Formulation tab) is used to initialize the following Target Solver component design parameter attributes:

Variables

The Lower and Upper bounds specified on input parameters in the general problem formulation are used as the Lower and Upper bounds for variables.

Targets

The targets specified for output parameters in the general problem formulation are used to initialize the value for corresponding selected targets.