The OPTEX Procedure

Summary of Functions

Table 14.3, Table 14.4, and Table 14.5 classify the OPTEX statements and options by function.

Table 14.3: Summary of Options for Specifying the Design

Function

Statement

Option

Design Characteristics

Number of design points

GENERATE

N=number

Saturated design

GENERATE

N=SATURATED

Augmented design

GENERATE

AUGMENT=SAS-data-set

Bayesian optimal design

MODEL

/ PRIOR=$p_1,p_2,\ldots $

Optimality Criteria

Minimize trace of $(X’X)^{-1}$

GENERATE

CRITERION=A

Maximize $|X’X|$

GENERATE

CRITERION=D

Minimize mean minimum

GENERATE

CRITERION=U

  distance to design

Maximize mean distance

GENERATE

CRITERION=S

  between nearest design points

Model Specification

Specify independent effects

MODEL

effects

Exclude intercept term

MODEL

effects NOINT

Specify CLASS variables

CLASS

variables

Specify CLASS variable parameterization

CLASS

/ PARAM=parameterization

Display CLASS variable parameterization

PROC OPTEX

CLASSPARAM

Static coding

PROC OPTEX

CODING=STATIC

Orthogonal coding

PROC OPTEX

CODING=ORTH

Orthogonal coding with

PROC OPTEX

CODING=ORTHCAN

  respect to candidates only

Suppress coding of effects

PROC OPTEX

NOCODE

Block Specification

Specify general covariance

BLOCKS

COVAR=SAS-data-set <options>

  matrix for runs

 

VAR=variables

Specify general covariate model

BLOCKS

DESIGN=SAS-data-set <options>

Specify b blocks of size k

BLOCKS

STRUCTURE=$(b)k$ <options>

Options for block specifications

  Repeat the search n times

 

ITER=n

  Retain best m searches

 

KEEP=m

  Select initial design at random

 

INIT=RANDOM

  Select initial design in order

 

INIT=CHAIN

Initial Design Characteristics

Random and sequential methods

GENERATE

INITDESIGN=PARTIAL$<(m)>$

Random initial design

GENERATE

INITDESIGN=RANDOM

Sequential initial design

GENERATE

INITDESIGN=SEQUENTIAL

Specify initial design

GENERATE

INITDESIGN=SAS-data-set


Table 14.4: Summary of Options for Searching for the Design

Function

Statement

Option

Design Search Specification

Retain best n searches

GENERATE

KEEP=n

Search n times

GENERATE

ITER=n

Specify candidate points

PROC OPTEX

DATA=SAS-data-set

Specify random seed

PROC OPTEX

SEED=number

Specify effective zero

PROC OPTEX

EPSILON=$\epsilon $

Design Search Methods

DETMAX algorithm with maximum excursion level

GENERATE

METHOD=DETMAX<(level)>

Exchange algorithm

GENERATE

METHOD=EXCHANGE

k-Exchange algorithm

GENERATE

METHOD=EXCHANGE$<(k)>$

Sequential algorithm

GENERATE

METHOD=SEQUENTIAL

Fedorov algorithm

GENERATE

METHOD=FEDOROV

Modified Fedorov algorithm

GENERATE

METHOD=M_FEDOROV


Table 14.5: Summary of Options for Examining and Saving the Design

Function

Statement

Option

Save the Design

Best design

OUTPUT OUT=SAS-data-set

 

Specific design

OUTPUT OUT=SAS-data-set

NUMBER=design-number

Block variable name

OUTPUT OUT=SAS-data-set

BLOCK=variable-name

Specify transfer variables

ID

variables

List the Design

Design characteristics

EXAMINE

 

Design points

EXAMINE

DESIGN

Information matrix $X’X$

EXAMINE

INFORMATION

Specific optimal design

EXAMINE

NUMBER=design-number

Variance matrix $(X’X)^{-1}$

EXAMINE

VARIANCE

Suppress all output

PROC OPTEX

NOPRINT