Yield optimization of cotton fiber, using a regression model based on the conditions of entry and process at the stage of ginning

Abstract

The yield and quality of cotton fiber obtained in the ginning process, determine the income for the farmer, the gin productivity and performance of the fiber in the process of industrialization. In this research, it was designed and programmed a hybrid metaheuristic between the simulated annealing (SA) and swarm particle (PSO) applied to a nonlinear regression model that predicts the performance of cotton fiber, made from cotton seed in ginning. To develop the model, input and process inherent to the conditions and factors to the 707 batches of seed cotton were analyzed during harvest ginning between 2009 and 2010, to which a total of 9 variables were considered (six and three binary integer), the development of this model allows to evaluate the expected performance and productivity set margins for farmers and for ginning enterprises. It was shown, that the seed cotton yield is adversely affected by such factors as: the variety of batch humidity as well as filth level, temperatura levels applied in clean fibers process.