A modeling and optimization approach to a tactical sugarcane harvest planning problem

In this paper, we address a tactical sugarcane harvest planning problem commonly faced by industries that cultivate, harvest, and process sugarcane in Brazil and some other countries. The problem involves a single mill that must harvest sugarcane from several geographically dispersed areas (harvest blocks) using capacitated mechanized harvest fronts. The objective is to minimize the total costs of the harvest process while satisfying operational constraints, including capacity limits for transportation and harvesting. We propose a new mixed-integer programming (MIP) model and develop MIP-based constructive and improvement heuristic procedures. In the constructive phase, the method generates a strong dual bound and an initial feasible solution, while the improvement phase refines this solution to achieve higher quality. The efficiency of the proposed methods is evaluated through computational experiments using a dataset based on real-world scenarios. Additionally, a sensitivity analysis provides managerial insights by examining different harvest front configurations.