基于FSVM的供应链节能减排危机预警研究

发布时间：2018-09-14 15:15

【摘要】：随着经济的快速发展,企业之间的竞争不断加剧,导致资源过度消耗,环境污染严重。在这种大背景下,我国提出“美丽中国”的构想,大力提倡节能减排,构建资源节约型与环境友好型社会。因此,企业必须在整个供应链内综合考虑产品的原材料获取、设计与生产、销售与运输、使用及回收再利用的过程中如何采用技术上可行、经济上合理以及环境与社会可承受的方式,以此达到能源消耗和污染物排放量最小、整体供应链环境负面影响最小以及资源利用率最高的目标。基于上述情况,供应链节能减排的提出就显得尤为迫切与重要。同时,供应链全球化趋势加剧,使得供应链内企业联系更加紧密,供应链结构更加复杂,任何一个环节出现问题都会对下一个环节的节能减排效果产生影响,进而出现供应链节能减排危机。所以,本文对供应链节能减排危机进行预警就显得十分必要。首先,本文从节能减排评价、供应链风险管理及危机预警模型这三个方面阐述了当今国内外的研究现状。简要介绍关于模糊C均值聚类(FCM)和支持向量机(SVM)的相关理论,为之后的研究做了扎实的理论准备。其次,借鉴财务危机的研究成果,依照明确警情一寻找警源一确定警素一预报警度一排警对策这五个步骤建立了供应链节能减排危机预警模块。其中,在寻找警源模块中,本文对供应链节能减排运作流程进行了较全面的分析,得到供应链节能减排危机警源分类体系。再者,在文献研究、政策解读和企业调研的基础上构建供应链节能减排危机预警初始警素体系,并通过专家调查法,让专家对各警素的可行性、重要性、区分性进行评价,得到专家的评价数据,最终形成了一套较为全面的供应链节能减排危机预警警素体系。最后,本文根据之前构建的供应链节能减排危机预警警素体系设计出企业调研问卷,并收集数据。然后应用模糊聚类(FCM)进行粗分,再利用模糊支持向量机(FSVM)对数据进行细分,给每个样本增加一个模糊隶属度,最终构建起基于FSVM的三级模糊支持向量机分类器。实证结果表明,基于FSVM的模型不仅提高了模型的分类速度及准确率,而且有效的解决了支持向量机的不可分问题,验证了模型的可行性。
[Abstract]:With the rapid development of economy, competition between enterprises is becoming more and more serious, which leads to excessive consumption of resources and serious environmental pollution. Under this background, our country puts forward the idea of "beautiful China", vigorously advocates energy saving and emission reduction, and constructs a resource-saving and environment-friendly society. Therefore, the enterprise must consider how the raw material acquisition, design and production, sale and transportation, use and recycling of products are technically feasible in the whole supply chain. In order to achieve the goals of minimum energy consumption and pollutant emission, minimum negative environmental impact of the whole supply chain and the highest utilization rate of resources, the economy is reasonable and the environment and society can afford it. Based on the above situation, the supply chain energy-saving and emission reduction is particularly urgent and important. At the same time, the trend of globalization of supply chain intensifies, which makes the enterprises in the supply chain more closely linked and the supply chain structure more complex. Any problem in any link will have an impact on the effect of energy saving and emission reduction in the next link. In turn, supply chain energy conservation and emission reduction crisis. Therefore, the supply chain energy-saving and emission reduction crisis warning is very necessary. First of all, this paper expounds the current research situation at home and abroad from three aspects: energy saving and emission reduction evaluation, supply chain risk management and crisis warning model. This paper briefly introduces the relevant theories of fuzzy C-means clustering (FCM) and support vector machine (SVM), and makes a solid theoretical preparation for the later research. Secondly, referring to the research results of financial crisis, according to the five steps of clear alarm situation to find out the police source and determine the alarm factor and pre-alarm degree and the alarm countermeasures, the paper establishes the supply chain energy-saving and emission reduction crisis warning module. Among them, in searching for the alarm source module, this paper has carried on the comprehensive analysis to the supply chain energy saving and emission reduction operation flow, obtains the supply chain energy saving and emission reduction crisis alarm source classification system. Furthermore, on the basis of literature research, policy interpretation and enterprise research, the paper constructs the initial warning system of supply chain energy saving and emission reduction crisis, and through expert investigation, let experts evaluate the feasibility, importance and distinction of each police element. Finally, a comprehensive warning system of energy saving and emission reduction crisis in supply chain is formed by obtaining the evaluation data of experts. Finally, according to the supply chain energy saving and emission reduction crisis warning system, the enterprise survey questionnaire is designed, and the data is collected. Then the fuzzy clustering (FCM) is used to divide the data, and then the fuzzy support vector machine (FSVM) is used to subdivide the data, adding a fuzzy membership degree to each sample. Finally, a three-level fuzzy support vector machine classifier based on FSVM is constructed. The empirical results show that the model based on FSVM not only improves the classification speed and accuracy of the model, but also effectively solves the inseparability problem of support vector machine, and verifies the feasibility of the model.
【学位授予单位】：福州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F274

【相似文献】