物联网技术环境下食品谱系建立与分析研究

Nowadays, the increasing food safety scandals attract people's concern very much. Most of the current precautions only emphasize inspections and tests by taking samples at the import, wholesale and retail levels. While, with the development of Cyber Physical Systems Engineering, it is a tendency that the application of wireless sensor network for food monitoring according the supply chain. Additionally, with the coming of globalization, there are also an increasing number of companies and organizations which have been involved into the food supply chain. It makes the management of food safety become more and more complex. This research aims to establish food pedigree model, to share information, to analyze record and finally to help making decisions based on a real-time monitoring system. Firstly, a experimental public web platform will be built for all the parties involved in the supply chain to share the information they have. Based on the information, the pedigree of products can be generated. It increases the transparency degree of the supply chain. The pedigree can be provided to the customers for a better understanding of the products they have bought. It can also boot the confidence in buying the product. Meanwhile, we will apply machine learning, data mining technology and social network theory to analyze the product pedigree. The results can help us to quantity product value and to find out the critical points and the weak points of the supply chain. More focus on the management and control of the points can effectively avoid food safety accidents. Once there have been foods contaminations, the system can also help to figure out where the problem lies through comparing the pedigree of all the problem products, to recall suspicious products for an active control and to lower the damages to a minimum level.

近年来不断爆发的食品污染事件引起人们对食品安全问题的高度关注。现有对食品污染的处理方法多重在事后处理而不是事前预防，往往造成巨大经济损失的同时还带来不好的社会影响。随着物联网技术的发展，对食品供应链进行实时检测成为可能，物联网技术也将成为未来保障食品安全的重要发展趋势。但在这一背景下，背后的食品质量管理系统还略显薄弱，食品供应链结构的复杂性也为研究带来挑战。本项研究旨在建立一个食品谱系模型为具有代表性的若干食品实现谱系可视化并对其进行分析。模型主要包括三部分，质量量化：将机器学习与食品质量变化模型结合建立一个动态的质量变化模型；谱系分析：以质量量化模型为标准，应用社会网络理论与量子算法对食品谱系进行分析，预防问题发生；危机响应：结合谱系分析，在问题发生时应用案例推理与数据挖掘技术，快速小代价找出问题原因并主动控制问题产品。该研究将为面向食品安全的管理系统提供了理论基础。

近年来不断爆发的食品污染事件引起人们对食品安全问题的高度关注。现有对食品污染的处理方法多重在事后处理而不是事前预防，往往造成巨大经济损失的同时还带来不好的社会影响。随着物联网技术的发展，对食品供应链进行实时检测成为可能，物联网技术也将成为未来保障食品安全的重要发展趋势。但在这一背景下，背后的食品质量管理系统还略显薄弱，食品供应链结构的复杂性也为研究带来挑战。本项研究开发了一个食品谱系分析与查询平台，为食品产品实现谱系可视化并对其进行分析。平台主要由三个模块构成，质量量化模块：将机器学习与食品质量变化模型结合建立一个动态的质量变化模型；谱系分析模块：以质量量化模型为标准，应用社会网络理论对食品谱系进行分析，预防问题发生；危机响应模块：结合谱系分析，在问题发生时应用数据挖掘技术，快速小代价找出问题原因并主动控制问题产品。在研究期间，主要成果为开发了一个食品谱系分析与查询平台，完成了9篇论文，其中6篇已经被录用，3篇仍在审阅中，申请了两项专利，其中一项已经获得授权。本项研究的科学意义主要在于1）使用ANN与k-NN混合算法对食品产品质量进行量化；2）对食品谱系关键环节进行分析：多个产品的谱系相互交织形成一个网络，网络中任何一个节点的变动也会对其点节点产生影响，本研究应用社会网络理论解决了该问题。3）本项研究重新设计关联规则算法，提高算法的搜索速度应用于解决危机处理与管理产品召回这一问题。