School of Industrial Engineering, College of Engineering, University of Tehran, Tehran, Iran.
Nord University Business School (HHN), Post Box 1490, 8049, Bodø, Norway.
Sci Rep. 2022 Mar 23;12(1):5037. doi: 10.1038/s41598-022-08993-5.
Industry 4.0 recommends a paradigm shift from traditional manufacturing to automated industrial practices, especially in different parts of supply chain management. Besides, the Sustainable Development Goal (SDG) 12 underscores the urgency of ensuring a sustainable supply chain with novel technologies including Artificial Intelligence to decrease food loss, which has the potential of mitigating food waste. These new technologies can increase productivity, especially in perishable products of the supply chain by reducing expenses, increasing the accuracy of operations, accelerating processes, and decreasing the carbon footprint of food. Artificial intelligence techniques such as deep learning can be utilized in various sections of meat supply chain management--where highly perishable products like spoiled meat need to be separated from wholesome ones to prevent cross-contamination with food-borne pathogens. Therefore, to automate this process and prevent meat spoilage and/or improve meat shelf life which is crucial to consumer meat preferences and sustainable consumption, a classification model was trained by the DCNN and PSO algorithms with 100% accuracy, which discerns wholesome meat from spoiled ones.
工业 4.0 建议从传统制造向自动化工业实践转变，特别是在供应链管理的不同环节。此外，可持续发展目标 12 强调了采用包括人工智能在内的新技术确保可持续供应链的紧迫性，以减少食物损失，这有可能缓解食物浪费。这些新技术可以提高生产力，特别是在通过降低成本、提高操作精度、加速流程和减少食物碳足迹来减少易腐产品供应链中的损失。深度学习等人工智能技术可用于肉类供应链管理的各个环节——在这些环节中，需要将像变质肉这样的高度易腐产品与健康产品分开，以防止与食源性病原体交叉污染。因此，为了实现这一过程的自动化，防止肉类变质和/或延长肉类保质期，这对消费者的肉类偏好和可持续消费至关重要，研究通过 DCNN 和 PSO 算法训练了一个分类模型，其准确率达到了 100%，可以区分健康肉和变质肉。
