Plant selection using a sequential decision algorithm based on an ideotype

Abstract

Objective: Develop a sequential decision algorithm to select plants in populations without repetitions based on an ideotype, using a logical process that employs everything from spreadsheets to programming languages.

Design/methodology/approach: A sequential algorithm based on conditional logic is proposed that implements a multi-criteria filtering process. The algorithm can be implemented on different numerical analysis platforms as it is based on a sequential logical decision process. To validate its application, a selection was made from a population of 60 coffee plants (Coffea arabica L.) of the Sarchimor T5296 variety, using Excel as the platform and the AND function due to its accessibility.

Results: Nine plants were selected that exhibited the desired characteristics of yield, plant height, canopy diameter, bienniality (I), and resistance to orange leaf rust, demonstrating the algorithmic selection capability.

Limitations on study/implications: The method has limitations inherent to its binary sequential design: (1) it does not differentially weight the importance of criteria, (2) it does not consider negative correlations between characters in simultaneous selection, and (3) it depends on thresholds defined a priori.

Findings/conclusions: The algorithm made it possible to select plants with behavior similar to or superior to the ideotype in contexts where conventional statistical methods are not applicable because they require replicated experimental designs.