Programming and scheduling sugarcane harvesting fronts: model and solution methods for large-scale problems

Abstract: In a recent study, optimization models were proposed for programming and scheduling sugarcane harvesting fronts. This is a complex agricultural and logistic problem comprising various factors, such as raw material maturation stage, harvesting at the agricultural unit, transporting of raw material to the plant, and milling capacities of the plant. In this study, one of the optimization models previously studied was used to represent this problem using Mixed Integer Programming (MIP) of a lot sizing and scheduling model in parallel machines with sequence dependent setup times and costs. The proposed methods are based on MIP heuristics to solve this model in a real situation of a harvest season of a typical company from this sector inspired by harvest block aggregation heuristics, relax-and-fix constructive heuristics, and fix-and-optimize improvement heuristics. To compare the performance of the heuristic methods, various experiments were conducted using different combinations and variations of these methods. Three approaches were able to produce good quality solutions. One of them is described in detail and analyzed in this study, showing promising results in terms of making programming and scheduling decisions concerning sugarcane harvesting fronts.

摘要：近期已有研究针对甘蔗收割作业编队的规划与调度问题提出了优化模型。该问题属于复杂的农业与物流耦合问题，涵盖原料成熟度、农业单元收割作业、原料向加工厂运输以及工厂研磨产能等多类影响因素。本研究采用此前已提出的一类优化模型，以混合整数规划（Mixed Integer Programming, MIP）框架对该问题进行建模，该模型为考虑序列相关调整时间与成本的并行机批量调度模型。本研究提出的求解方法基于混合整数规划启发式算法，针对该行业某典型企业的真实收割季场景设计，算法框架借鉴了收割区块聚合启发式、松弛-固定构造启发式与固定-优化改进启发式三类策略。为对比各启发式方法的求解性能，本研究基于上述启发式的不同组合与变体设计了多组对比实验，最终有三类方法可生成高质量的可行解。本研究对其中一类方法进行了详细阐述与性能分析，结果显示该方法在甘蔗收割作业编队的规划与调度决策中表现出良好的应用前景。