This paper addresses a multi-objective sugarcane harvesting problem in Thailand, where several conflicting objectives and local restrictions are regarded as major obstacles to a sustainable sugar production environment. A multi-objective modeling approach that balances three different objectives of different key supply chain actors, namely (i) maximizing output in terms of total sugar production volume, (ii) maximizing grower equity in terms of a fair harvesting time-slot distribution, and (iii) maximizing supply chain efficiency in terms of a lower variability in resource requirements across the harvesting season, is introduced and solved by a state-of-the-art multi-objective evolutionary genetic algorithm. To better help the algorithm generate efficient solutions forming the Pareto front, two local searches are also embedded and intermittently performed during algorithm execution. Based on the information of an operating mill in Kanchanaburi Province, Thailand, we have found that our approach produces solutions that are close to optimal in terms of production output. Nonetheless, by sacrificing a small amount of production output, these solutions provide significant improvements in terms of grower equity and supply chain resource efficiency, which are crucial for the survivability of involved actors.

本文解决了泰国的多目标甘蔗收获问题，该国几个相互矛盾的目标和当地限制被视为可持续糖业生产环境的主要障碍。一种多目标建模方法，可以平衡不同主要供应链参与者的三个不同目标，即（i）以总糖产量为单位使产量最大化，（ii）以公平的收获时间段分配为基础使种植者权益最大化，以及（iii）通过最先进的多目标进化遗传算法引入并解决了最大的问题，即在整个收获季节减少资源需求的可变性来最大化供应链效率。为了更好地帮助算法生成形成Pareto前沿的有效解，还嵌入了两个局部搜索，并在算法执行过程中间歇执行。根据泰国北碧府一家工厂的信息，我们发现我们的方法所产生的解决方案在产量方面接近最佳。但是，通过牺牲少量生产产量，这些解决方案在种植者权益和供应链资源效率方面提供了显着改善，这对于相关参与者的生存能力至关重要。