The OPTEX Procedure


PROC OPTEX Statement

  • PROC OPTEX <options>;

You use the PROC OPTEX statement to invoke the procedure. The following options can be used:

CLASSPARAM

specifies that a table should be displayed summarizing the parameterization of classification variables in the model for the design.

CODING=NONE
CODING=STATIC
CODING=ORTH
CODING=ORTHCAN

specifies which type of coding to use for modeling effects in the design. Coding equalizes all model effects as far as the optimization is concerned. The default is CODING=STATIC, which specifies that the values of all effects are to be coded to have maximum and minimum values of +1 and –1, respectively. The options CODING=ORTH and CODING=ORTHCAN specify orthogonal coding with respect to the points in the candidate data set. The option CODING=NONE suppresses coding of effects; it is equivalent to the NOCODE option. For more details on coding, see the section Design Coding.

Note that while CODING=STATIC is the default, CODING=ORTH will usually give more meaningful efficiency values, especially if all possible combinations of factor levels occur in the candidate data set.

DATA=SAS-data-set

specifies the input SAS data set that contains the candidate points for the design. By default, the OPTEX procedure uses the most recently created SAS data set. For details, see the section DATA= Data Set.

EPSILON=$\epsilon $

specifies the smallest value $\epsilon $ that is considered to be nonzero for determining when the search is no longer yielding an improved design and when the information matrix for the design is singular. By default, $\epsilon $ = 0.00001.

NAMELEN=n

specifies the length of effect names in tables and output data sets to be n characters long, where n is a value between 20 and 200 characters. The default length is 20 characters.

NOCODE

suppresses the coding of effects in the model for the design. This option is equivalent to CODING=NONE.

NOPRINT

suppresses all output. This is useful when you only want the final design to be saved in a data set.

SEED=s

specifies an integer used to start the pseudo-random number generator for initialization (see the section Search Methods). If you do not specify a seed, or if you specify a value less than or equal to zero, the seed is by default generated from reading the time of day from the computer’s clock.

STATUS=status-level

specifies that the status of the search be checked at the given level, where status-level is an integer between 1 and 4, inclusive. If you specify a status-level, then a table of the status at each check point is displayed. You can use this table to track the progress of long searches. The allowable status-levels are listed in the following table:

Status-level

Checks status after each:

1

design search; the number of searches specified by the NITER= option

2

search loop

3

internal search loop

4

extra internal search loop for METHOD=M_FEDOROV

Each search method loops to produce successively better designs; these are the search loops for STATUS=2. STATUS=3 and STATUS=4 refer to deeper loops within the search methods. You will only need to specify STATUS=3 or STATUS=4 very rarely, since evaluating and displaying the status at either of these levels will usually make the search much slower.