Improved differential evolution algorithms for solving multi-stage crop planning model in southern region of Thailand.

This paper presents algorithms based on Differential Evolution and Improved Differential Evolution for solving a multi-stage crop planning problem in southern region of Thailand to maximize the profit. Four types of algorithms were tested: 1) Differential Evolution (DE), 2) Differential Evolution with local search by adding the step of local search after the selection process, which used insert algorithm (DE-I), 3) Random best of Differential Evolution improved by mutations (DE-R), 4) Random best of Differential Evolution with local search as a mixture of types 2 and 3 (DE-IR). The results show that with small problem instances all the algorithms found a 100% optimal solution. In medium and large problem instances DE-IR shown the best solutions among the proposed algorithms. [ABSTRACT FROM AUTHOR]