基于改进萤火虫优化LSSVM算法的碳排放影响因素研究

发布时间：2018-03-25 14:03

本文选题：碳排放影响因素　切入点：最小二乘支持向量机　出处：《华北电力大学》2017年硕士论文

【摘要】：能源消费一直是人类经济发展和社会进步的重要驱动力,特别是对中国来说,煤炭和原油等化石燃料主导了国家人民的生产生活。但是,在全球气候变化日益严峻以及低碳可持续发展经济兴起的背景下,能源消费过程中随之带来的二氧化碳排放问题已经不可忽视。作为第一的发展中国家,中国如何在保证国内经济又好又快发展的同时,确保能源安全供给、减少碳排放量、应对气候变化以及承担国际环境保护义务与责任,成为亟待解决的一大问题。河北省作为中国大省,近些年来,其环境问题极其严重,特别是雾霾污染问题。考虑到碳排放现象是引起雾霾的一大重要因素,对河北省影响碳排放因素的研究是十分必要的。在本文中,首先介绍了与碳排放相关的概念与理论,中国碳排放现状以及目前国内外对于碳排放研究的具体情况。其次,本文介绍了一种适合小样本数据的最小二乘支持向量机算法,针对算法惩罚因子和核函数宽度根据经验确定的问题,本文引入改进的萤火虫算法,利用启发式智能算法进行寻优,在此基础上提高算法的精确度。在实证研究部分,本文选择了1990-2014年中国河北省的碳排放量和影响因子相关数据为研究对象。对于预选的多个影响因子,本文采用SPSS统计软件确定因子的显著相关性。然后根据样本数据的特点,运用本论文提出的算法针对性进行建模。为了更好地定量衡量因子对碳排放量的具体影响,后续还采用了可拓展的随机性的环境影响评估模型(STIRPAT)与对数平均迪氏分解模型(LMDI)深入分析碳排放量与影响因子之间的关联性。根据分析结果,可以得知:(1)与其他三种算法对比,本文提出的新算法验证了河北省碳排放量与经过SPSS筛选确定的十三个影响因子之间的因果关系,证明了本文提出的算法对标准萤火虫算法的改进和对最小二乘支持向量机的优化是有效的。(2)STIRPAT模型结果表明河北省碳排放量与本文选定的13个影响因子之间均为正向相关,其中最终消费的驱动指数最大,交通运输工具拥有量的指数最小。(3)利用LMDI分解法对碳排放系数、能源强度、能源消费结构、产业结构、经济活动规模和人口规模这六个影响因素进行分解,得知经济活动规模效应是所有因素中对河北省碳排放总量影响最大的正向推动因素。最后,本文还从能源结构、能源效率、人口政策、交通工具和低碳理念层面提出了一些针对性的建议,有利于在理论层面上为政府制定减排政策提供支持,并从源头上有效控制碳排放的产生。
[Abstract]:Energy consumption has been an important driving force for human economic development and social progress, especially for China, where fossil fuels such as coal and crude oil dominate the production and livelihood of the country's people. In the context of the increasingly severe global climate change and the rise of low-carbon sustainable development economy, the carbon dioxide emissions brought about by the energy consumption process can not be ignored. As the first developing country, How can China ensure a sound and rapid development of its domestic economy while ensuring a secure supply of energy, reducing carbon emissions, combating climate change and assuming international environmental protection obligations and responsibilities? Hebei Province, as a big province in China, has been facing serious environmental problems in recent years, especially haze pollution. Considering that carbon emission is one of the most important factors causing haze, It is necessary to study the factors affecting carbon emissions in Hebei Province. In this paper, the concepts and theories related to carbon emissions, the current situation of carbon emissions in China and the specific situation of carbon emission research at home and abroad are introduced. This paper introduces a least squares support vector machine (LS-SVM) algorithm for small sample data. Aiming at the problem that penalty factor and kernel width are determined by experience, an improved firefly algorithm is introduced in this paper. Heuristic intelligent algorithm is used to optimize and improve the accuracy of the algorithm. This paper selects the data of carbon emissions and impact factors from 1990 to 2014 in Hebei Province, China as the research object. For the pre-selected factors, we use the SPSS software to determine the significant correlation of the factors. Then according to the characteristics of the sample data, In order to better measure the specific impact of factors on carbon emissions, In the follow-up, the extended stochastic environmental impact assessment model (STIRPAT) and the logarithmic average Dickers decomposition model (LMDI) were used to analyze the correlation between carbon emissions and impact factors in depth. According to the results of the analysis, we can see that: 1) is compared with the other three algorithms. The new algorithm proposed in this paper verifies the causal relationship between the carbon emissions in Hebei Province and the thirteen influence factors determined by SPSS screening. It is proved that the improvement of the standard firefly algorithm and the optimization of the least square support vector machine (LS-SVM) are effective. The results show that the carbon emissions of Hebei Province are positively correlated with the 13 factors selected in this paper. Among them, the driving index of final consumption is the largest, the index of ownership of transportation vehicles is the smallest. (3) the carbon emission coefficient, energy intensity, energy consumption structure and industrial structure are analyzed by LMDI decomposition method. By decomposing these six factors, the scale of economic activity and the size of population, we know that the scale effect of economic activity is the most positive driving factor to the total amount of carbon emissions in Hebei Province. Finally, this paper also analyzes the energy structure and energy efficiency. Population policy, transportation and low-carbon concept level put forward some targeted suggestions, which is conducive to the theoretical support for the government to formulate emission reduction policies, and to effectively control the generation of carbon emissions from the source.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X24;F426.2

【参考文献】