废润滑油再生决策支持系统的研究

针对废润滑油再生过程中的工艺选择,将决策支持系统引入废润滑油再生领域。提出了废润滑油再生工艺决策支持系统的理论框架。构建出基于数据库、模型库的废润滑油再生工艺选择决策系统。用户可通过良好的人机对话界面,根据润滑油污染程度和再生要求,确定满足需要的废润滑油再生工艺。     

小波支持向量机在大气污染物浓度预测中的应用

用小波分解重构和支持向量机相结合的方法,建立大气污染物浓度预测模型。通过小波分解,将大气污染物浓度序列分解为不同频段的小波系数序列,再对各层的小波系数序列重构到原尺度上。利用相关分析的方法构建出低频小波系数a3和中频小波系数d3的支持向量机模型输入因子为前一天小波系数a3和7个气象因子;高频小波系数d2和d1以前三日的小波系数为输入因子,然后对各小波系数序列采用相应的支持向量机模型进行预测,各小波系数均使用ν-支持向量回归机(ν-SVR)算法和径向基函数,最后通过小波重构合成大气污染物浓度序列的最终预测结果。通过对大气SO2浓度预测实例证明,该大气污染物浓度预测模型具有推广能力较强、预测精度较高、训练速度快、便于建模等优点,具有良好的应用前景。     

基于空间数据挖掘的区域危险废物决策支持设计研究

区域危险废物环境管理与风险控制是涉及到危险废物特性、环境、经济和社会等多因素的复杂过程,对资源环境的可持续发展和人体健康都至关重要。空间数据挖掘技术在传统数据挖掘的基础上增加空间维尺度,能够实现从空间数据库和非空间数据库中提取危险废物产生、转移、处置和管理过程中隐含的空间模式与特征。文章在传统的危险废物环境管理和风险控制方法的基础上,构建了区域危险废物空间数据挖掘的指标体系,并设计了一种基于空间数据挖掘技术的区域危险废物决策支持系统框架。该框架由数据层、数据挖掘层和显示层组成,为区域危险废物管理和决策提供了直观的、科学的和高效的支持,提供了一种新颖可行的思路和方法提高危险废物的环境管理和风险控制水平,具有一定的理论和现实意义。     

Factors affecting the evacuation decisions of coastal households during Cyclone Aila in Bangladesh

Despite sincere efforts by concerned agencies and recent improvements in hazard warnings, thousands of at-risk people did not evacuate during Aila, a category-I tropical cyclone that struck southwestern coastal Bangladesh in 2009. This study investigated the responses of the people affected by Aila to cyclone warnings and associated evacuation orders, and unveiled behavioural aspects that explain why they did or did not comply with the evacuation orders. Utilising the primary data collected from 420 households living in the severely affected coastal sub-district Koyra, located in Khulna District, this study found that although more than 97% of the households had received cyclone warnings and evacuation orders before Aila's landfall, only around 26% had evacuated. We also examined this study's empirical findings for factors that had dissuaded people from complying with the evacuation orders. Relevant test statistics along with results from principal component analysis suggested that the significant and systematic absenteeism of households from disaster preparedness training appears to be one of the key determinants that affected their evacuation decisions, along with factors related to warning messages, the attributes of cyclone shelters, risk perception, and socioeconomic aspects. A number of recommendations are also presented in this study for people at risk to improve their evacuation rates in future emergencies, not only to save their own lives but also to save their livestock.     

考虑颜色特征最优组合的CART决策树火灾图像识别方法*

针对火灾图像识别过程中颜色特征数量多、特征间相关性复杂、难以在多维特征融合过程中有效融合图像颜色特征等问题,提出1种考虑颜色特征最优组合的CART决策树火灾图像识别方法。首先,在Lab、RGB、HSV 3种色彩模式下基于图像颜色特征提取火灾图像特征序列;其次,分别在3种色彩模式下基于精细决策树与特征随机排列组合方法提取颜色特征中最优组合特征;最后,将提取的火灾图像最优组合特征序列作为CART决策树输入进行模型训练,并通过测试样本以及其他机器学习方法进行模型泛化能力的分析。研究结果表明:本文方法寻找出识别火灾图像的最优颜色特征组合为“Kb1+Var1+Kg+Kb2+Var2+Kh+Ks+Kv”;CART决策树方法对于火灾图像识别的测试准确度可达84.5%,其分类效果明显优于其他决策树类与集成树类方法;9折为最佳交叉验证折数,其测试准确度可达86.47%,与5折交叉验证相比明显提升14.77%。研究结果可为火灾图像识别提供方法基础。     

基于熵值法和支持向量机的煤矿应急救援能力评价模型

为增强煤矿应急救援能力,提高评价的精度,弥补单一评价方法的缺陷,构建了基于熵值法和支持向量机的煤矿应急救援能力评价模型。根据各评价指标数据所提供的信息量,运用熵值法设定指标权重,进一步将指标权重定量化、客观化;同时鉴于煤矿应急救援能力评价数据难收集、影响因素复杂多变的状况,结合具有处理非线性、小样本数据问题的支持向量机进行评价,保证了评价结果的准确性。实证分析表明,该模型的评价结果与煤矿实际应急救援能力非常接近,可用于实际的煤矿应急救援能力评价。     

软岩大变形巷道刚柔结合支护方法研究

深部软岩地下工程围岩稳定性差且具有长期流变效应,由于支护不利导致围岩产生大变形破坏问题趋于突出。针对以上问题,以某深部软岩巷道工程建设为例,采用现场监测、数值模拟等方法,分析围岩长期变形趋势及破坏机理;针对围岩破坏特征,提出“刚柔”结合的软弱大变形巷道围岩新型支护方法:考虑围岩不同深度破坏情况不同,采用不同注浆方法分区域加固,形成“刚性”承载圈,并在巷道最外部设置缓冲层及可缩性U型钢支架,形成柔性加固圈。利用有限元模型对其支护效果进行数值模拟,结合现场监测变形数据分析,验证了新型支护方案的合理性。     

农户有限理性土地利用行为决策影响因素——以陕西省米脂县高西沟村为例

利用有限理性理论与方法进行行为决策研究已成为多样化决策研究中的热点。研究微观尺度土地利用变化中农户有限理性行为决策的影响因素,对于揭示农户土地利用驱动因子、调整农户决策行为与优化农业景观格局具有重要的理论与现实意义。论文以陕西省米脂县高西沟村为例,通过构建农户效应收益模型,利用2009—2011年农户调查数据揭示理性最优决策与农户实际决策之间的差异性;利用体现农户行为满意度的Probit模型分析有限理性决策的影响因素。结论为:①高西沟村农户理性最优决策与实际决策差异性较大,说明农户在实践中采用有限理性进行土地利用生产;②构建Probit模型模拟农户满意度决策行为态度,选取的4类变量较好地解释了农户决策的影响因子。其中,农业收入回归系数为0.78、反映相对经济地位下降变量0.41、退耕面积-0.42、耕地面积-0.62、打工收入-0.34、性别0.48、劳动力0.31,均处于显著相关水平。各变量估计结果与理论预测和实际调查结果相符合,具有较强的解释力。     

Interaction Between Scientists and Nonscientists in Community-Based Watershed Management: Emergence of the Concept of Stream Naturalization

The authors' personal experience in watershed planning and decision making in the agricultural Midwest is described to illustrate how: (1) formalization of the process of community-based management is not sufficient to guarantee that local people will meaningfully consider scientific information and opinion when making decisions about watersheds, and (2) genuine social interaction between scientists and nonscientists requires a considerable investment of time and energy on the part of the scientist to develop personal relationships with nonscientists based on trust and mutual exchange of information. This experience provides the basis for developing a general conceptual model of the interaction between scientists and nonscientists in community-based watershed management in the agricultural Midwest. An important aspect of integrating science effectively into community-based decision making is the need to revise existing concepts to accommodate place-based contexts. Stream naturalization is introduced as an alternative to stream restoration and rehabilitation, which are viewed as inappropriate management strategies in human-dominated environments. Stream naturalization seeks to establish sustainable, morphologically and hydraulically varied, yet dynamically stable fluvial systems that are capable of supporting healthy, biologically diverse aquatic ecosystems. This general goal is consistent with the types of stream-management practices emerging from community-based decision making in human-dominated, agricultural landscapes. Further research on the linkages between geomorphological and ecological dynamics of human-modified agricultural streams over multiple spatial and temporal scales is needed to provide a sound scientific framework for stream naturalization.