Authors: Angelo Aliano Filho; Helenice De Oliveira Florentino; Margarida Vaz Pato
Addresses: Curso de Biometria, Departamento de Bioestatística, IB, UNESP, 18618-970, Botucatu, SP, Brazil ' Departamento de Bioestatística, IB, UNESP, 18618-970, Botucatu, SP, Brazil ' CIO and ISEG, Universidade de Lisboa, 1200-781, Lisboa, Portugal
Abstract: This paper presents a mathematical model adapted from literature for the crop rotation problem with demand constraints (CRP-D). The main aim of the present work is to study metaheuristics and their performance in a real context. The proposed algorithms for solution of the CRP-D are a genetic algorithm, a simulated annealing and hybrid approaches: a genetic algorithm with simulated annealing and a genetic algorithm with local search algorithm. A new constructive heuristic was also developed to provide initial solutions for the metaheuristics. Computational experiments were performed using a real planting area and semi-randomly generated instances created by varying the number, positions and dimensions of the lots. The computational results showed that these algorithms determined good feasible solutions in a short computing time as compared with the time spent to get optimal solutions, thus proving their efficacy for dealing with this practical application of the CRP-D.
Keywords: optimisation; metaheuristics; crop rotation; mathematical modelling; genetic algorithms; simulated annealing.
International Journal of Metaheuristics, 2014 Vol.3 No.3, pp.199 - 222
Available online: 15 Oct 2014 *Full-text access for editors Access for subscribers Purchase this article Comment on this article