A bi-objective genetic approach for the selection of sugarcane varieties to comply with environmental and economic requirements

Abstract

The selection of sugarcane varieties is an important problem faced by companies in Brazil that exploit sugarcane harvest for energy production. In the light of current concerns regarding the reduction of environmental damage and the efficiency of the production system, research into this problem is called for. In this context the authors begin by outlining the sugarcane variety selection problem in accordance with technical constraints with the purpose of minimizing collection and transport costs and maximizing energy balance obtained from residues of the sugarcane harvest. They then present a previously developed model for the problem within bi-objective binary linear programming and study its computational complexity. Fundamentally, this paper is devoted to the application of a bi-objective genetic heuristic to the question addressed. A computational experiment, performed by resorting to a test set including real and semi-randomly generated instances, is then reported. The results prove the high quality of the heuristic in terms of solution quality, besides computing time. For these reasons, this will be an appropriate tool to help sugarcane company managers to plan their producing activities.

Keywords:

• agriculture
• costing
• energy
• integer programming
• greedy heuristics
• bi-objective genetic heuristics

Acknowledgements

We thank FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo, Grant No. 2009/14901-4 and No. 2010/07585-6, Brazil), FUNDUNESP (Fundação para o Desenvolvimento da UNESP, Brazil) and PROPG UNESP for their financial support. The research of the second author was also partially funded by FCT (Fundação para a Ciência e Tecnologia, Portugal) under the project POCTI/ISFL/152.