Output

By default, the OPTEX procedure lists the following information for each attempt to find the optimum design:

  • the D-efficiency of the design

  • the A-efficiency of the design

  • the G-efficiency of the design

  • the square root of the average variance for prediction over the candidate points

If you specify a BLOCKS statement, then the covariate-adjusted D- and A-efficiencies are also listed.

See the section Design Efficiency Measures for details on the efficiencies. The OPTEX procedure orders the designs first by the optimality criteria with which they were generated and then by optimality with respect to the other three preceding measures.

If you use the NOCODE option, the OPTEX procedure lists

  • the G-efficiency of the design

  • the square root of the average variance for prediction over the candidate points

If you specify one of the distance-based optimality criteria (CRITERION=U or CRITERION=S), then, instead of the preceding efficiencies, alternative measures of coverage and spread are listed. For U-optimality these measures are

  • the average distance from each candidate to the nearest design point (this is the U criterion)

  • the average harmonic mean distance from each candidate to the design

For S-optimality, the following alternative measures of spread are listed:

  • the harmonic mean distance from each design point to the nearest other design point (this is the S criterion)

  • the average distance from each design point to the nearest other design point

In addition, the OPTEX procedure can create an output data set, as described in the section OUTPUT Statement and in the section Output Data Sets.