Inefficiencies in the food supply chain account for up to 60% of food wasted in the United States, significantly inhibiting efforts to tackle food insecurity. In this work, this problem is addressed by developing a supply chain decision-making framework that explicitly considers complex biochemical product quality degradation processes as a function of environmental conditions (e.g., temperature, humidity, atmospheric composition). The resulting optimization problem is solved online in real-time to mitigate demand uncertainty, reducing operating costs, and inventory spoilage. We demonstrate that this approach is equivalent to a data-driven, feedback-based control strategy that relies on manipulating environmental conditions at storage facilities and in transportation equipment. Since large-scale supply chain network instances result in computationally prohibitive optimization problems, a novel and highly efficient heuristic is introduced, that allows for obtaining solutions in practical amounts of time and with negligible degradation in the value of the objective function. The performance of our proposed approach is benchmarked with extensive numerical simulations based on a realistic, large-scale study of the produce supply chain from Mexico to the United States.

食品供应链的低效率占美国浪费的食物的 60%，严重阻碍了解决粮食不安全问题的努力。在这项工作中，这个问题是通过开发一个供应链决策框架来解决的，该框架明确地将复杂的生化产品质量降解过程视为环境条件（例如，温度、湿度、大气成分）的函数。由此产生的优化问题在线实时解决，以减轻需求不确定性、降低运营成本和库存损坏。我们证明这种方法相当于数据驱动、基于反馈的控制策略，它依赖于操纵存储设施和运输设备的环境条件。由于大规模供应链网络实例导致计算上禁止优化问题，因此引入了一种新颖且高效的启发式方法，它允许在实际时间内获得解决方案，并且目标函数值的下降可忽略不计。我们提出的方法的性能以基于对从墨西哥到美国的农产品供应链的现实、大规模研究的广泛数值模拟为基准。