A new method for optimal cropping pattern

Abstract

This work proposes the GenSRT method for the Cropping Pattern Optimization (CPO) problem. GenSRT applies Genetic Algorithms, Simplex Method, and Regression Trees. The purpose is to maximize the net income of every cropping season. Simplex Method (SM) is the traditional approach for solving the problem; however, CPO is complicated, because the crop yield has a non-linear behavior which SM cannot consider directly. In GenSRT, regression trees are applied to non-linear regression models construction. The models are provided to GenSRT to evolve efficient cropping patterns and to maximize the benefits for sowing area distribution through a genetic search. Results show that GenSRT overcomes Simplex maximization by obtaining better resource distribution and a higher net profit. © 2009 Springer-Verlag Berlin Heidelberg.