Generating designs for comparative experiments with two blocking factors

Area of Science:

• Agricultural Science
• Experimental Design
• Statistical Modeling

Background:

• Comparative experiments often feature one treatment factor and two blocking factors, common in agricultural research.
• Designing experiments with numerous treatments and complex blocking structures presents significant challenges.
• Existing methods for generating optimal designs can be computationally intensive and limited.

Purpose of the Study:

• To develop a novel search algorithm for efficiently finding optimal experimental designs with two blocking factors.
• To compare the performance of the new algorithm against established software (CycDesigN, DiGGer, SAS OPTEX).
• To demonstrate the algorithm's applicability and efficiency across various experimental designs.

Main Methods:

• A new search algorithm incorporating efficient update formulas was developed.
• Augmented row-column designs were generated and compared with existing methods.
• The algorithm was applied to generate two-phase and incomplete row-column designs.

Main Results:

• The new algorithm significantly reduces computing time for optimal design generation.
• It provides high-quality solutions, comparable or superior to existing methods.
• The efficient update formulas offer advantages in specific scenarios where existing formulas fail.

Conclusions:

• The proposed algorithm offers an efficient and effective approach for generating optimal experimental designs with two blocking factors.
• This method is particularly beneficial for complex agricultural experiments with large treatment numbers.
• The framework supports the generation of augmented row-column, two-phase, and incomplete row-column designs, broadening its practical utility.